Screening for Autochthonous Phytoextractors in a Heavy Metal Contaminated Coal Mining Area

In order to protect public health and crops from soil heavy metal (HM) contamination at a coal mining area in Henan, central China, HM pollution investigation and screening of autochthonous HM phytoextractors were conducted. The concentrations of cadmium (Cd), lead (Pb), copper (Cu) and zinc (Zn) in surface soils exceeded the corresponding local background values and the China National Standard (CNS). The maximum potential ecological risk (RI) was 627.30, indicating very high ecological risk. The monomial risk of Cd contributed the most to the RI, varying from 85.48% to 96.48%. The plant community structure in the study area was simple, and was composed of 24 families, 37 genera and 40 species. B. pilosa, A. roxburghiana, A. argyi, A. hispidus were found to be the most dominant species at considerable risk sites. Based on the comprehensive analysis of Cd concentration, bioconcentration factor, translocation factor and adaptability factor, B. pilosa and A. argyi had potential for phytoextraction at considerable risk sites. A. roxburghiana had potential for Cd phytoextraction at moderately risk sites and A. hispidus seemed suitable for phytostabilization. The results could contribute to the phytoremediation of the similar sites.


Introduction
The coal industry is one of the most important pillars of China's economy. The coal mining activity produces a large amount of coal gangues, from which toxic HMs could be released during the destruction of gangue mineral structure under the combined effects of water, microorganisms, vegetation, sunlight radiation and heat [1]. The HMs, unlike organic pollutants, cannot be biodegraded or biodeteriorated. As a consequence, they accumulate in the environment. In addition, HMs from the environment also migrate toward and accumulate in the living organisms by the processes of bioaccumulation and biomagnification through the trophic levels of the ecosystem. The accumulation of HMs in agricultural soils and water bodies poses a considerable threat to human health by different exposure pathways-direct ingestion, dermal absorption, inhalation and food chains [2,3]. To deal with these problems, a variety of physical and chemical treatment approaches have been developed, such as in situ vitrification, soil incineration, excavation and landfill, soil washing, soil flushing, solidification, and stabilization of electrokinetic systems [4,5]. However, these techniques have certain drawbacks like high expense, intensive labor, secondary pollution and irreversible change of the soil properties [6]. 2 of 14 As one of the novel remediation methods, phytoremediation is considered as a cost-effective and environmentally-friendly solution to the HM pollution problem.
Phytoremediation refers to the use of vegetation and associated soil microbes to lower the concentrations or hazardous effects of contaminants in the environments [7]. Generally, phytoremediation includes five techniques: phytoextraction (phytoaccumulation), phytofiltration, phytostabilization, phytovolatilization, and phytodegradation [8]. Phytoextraction is the HM absorbtion by plant roots, translocation and accumulation in above-ground (AG) biomass [9,10], which is the most useful phytoremediation technique for the removal of HMs from contaminated sediments or water [11,12]. In the process of phytoextractor selection, the native plants should be taken as a priority. The reasons are from the following points: (a) they tend to have higher viability [13]; (b) the autochthonous plants growing at contaminated sites are also considered as a valuable source of seed banks and gene pools [14]; (c) most importantly, alien plant species could reduce local plant species diversity [15][16][17] and pose non-negligible impacts at the local species, community and ecosystem level [18]. Many studies of HM phytoremediation with the native flora have been conducted [19,20].
The bioconcentration factor (BCF) and translocation factor (TF) are the most utilized indices for phytoremediation plant evaluation and selection [21]. However, they are not able to thoroughly characterize the phytoremediation ability of in situ plant species, because the phytoextraction efficiency depends on not only soil HM bioavailability and soil properties, but also the characteristics of the plant species like growth rate, AG biomass production, tolerance to the target HM toxicity, and adaptability to local environmental conditions [22][23][24]. Therefore, adaptability factor was introduced as an amendment of the screening criterion in this study. A typical coal gangue dump and the adjacent land in Jiaozuo, central China, were chosen as the study area. A previous study conducted at this same site showed that the gangue dump produced HM pollution in the adjacent cultivated land controlled by prevailing wind and posed a considerable threat to the environment [25]. In order to ensure the environmental safety during the reuse of the dump site and the crop safety of the surrounding farmland require systemic remediation and restoration. This study aimed to: (a) investigate and assess HM pollution and screen of target HM for phytoremediation. The metals evaluated are Cd, Pb, Cu, Cr and Zn; (b) survey local vegetation and determine the HM contents in different native species; (c) analyze the corresponding adaptability factors, bioconcentration factors and translocation factors in order to screen out suitable native plant species for phytoremediation of the target HM.

Site Description
The study area (N 35 • 16 2.1 , E 113 • 21 8.2 ) is located in Yanmazhuang (Jiaozuo, central China, Figure 1). The elevation of the terrain gradually decreases from the northwest to the southeast; the direction of the groundwater flow is to the southeast (130 • ). The area is a part of the front Taihang piedmont alluvial-pluvial fan, which is under the control of a temperate continental monsoon climate. There are three types of soil: sandy loam, clay loam and silty clay loam. The groundwater depth varies between 1 m and 3 m. The prevailing wind directions are northeast and southwest. The annual average temperature, precipitation and average evaporation are 14 • C, 610 mm and 2039 mm, respectively. The main crops are summer maize (Zea mays) and winter wheat (Triticum aestivum).

Sample Collection and Analysis
The investigation was mainly carried out at the coal gangue dump site and the adjacent land, which covered around 30 km 2 . All sampling sites were chosen with caution to avoid anthropogenic disturbances such as the main roads, railways, etc. Thirteen surface soil/gangue and twenty-one plant samples were collected. The soil samples were collected at the depth of 3-10 cm with wooden chips. The sampling density varied from 50 to 500 m according their distance from the gangue dump. A higher sampling density was chosen closer to the dump. Each sample was collected in triplicate.
The sampling density varied from 50 to 500 m according their distance from the gangue dump. A higher sampling density was chosen closer to the dump. Each sample was collected in triplicate. Each duplicate plant sample was formed with 3-5 different plant individuals and each duplicate soil sample was the mixture of the soil subsamples collected beneath the sampled plant individuals. Ten additional plant samples were also collected. Since these species were unlikely phytoextractors according to the literature review, duplicate sampling was not applied. The plant samples were separated into the above-ground part (shoots) and root parts. Representative samples for each species were selected for the herbaria and subsequent identification. GPS coordinates of all sampling sites were documented. Soil pH was measured by an IQ150 portable pH-conductivity meter (Spectrum Technologies, Inc., Aurora, IL, USA).
The plant samples were washed three times using distilled water and oven-dried. The samples were ground, sieved through 100 mesh sieve, and digested with an H2O2-HClO4-HF-HNO3 method in PTFE jars on an electric heating plate. Two Chinese national standard samples (GBW07427, GBW10048), as well as a blank sample was used as references to control the quality of the sample analysis. The total concentrations of cadmium (Cd), lead (Pb), copper (Cu), chromium (Cr), and zinc (Zn) in plants, soil and gangue were measured by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, IRIS Intrepid II XSP, Thermo Scientific, Waltham, MA, USA). The recovery rates of the standard samples were within 90-110%.

Potential Ecological Risk Assessment
The concept of potential ecological risk assessment (RI) was introduced by Hakanson in 1980, and is widely utilized in soil and sediment HM studies. The calculation formulas are as follows: Each duplicate plant sample was formed with 3-5 different plant individuals and each duplicate soil sample was the mixture of the soil subsamples collected beneath the sampled plant individuals. Ten additional plant samples were also collected. Since these species were unlikely phytoextractors according to the literature review, duplicate sampling was not applied. The plant samples were separated into the above-ground part (shoots) and root parts. Representative samples for each species were selected for the herbaria and subsequent identification. GPS coordinates of all sampling sites were documented. Soil pH was measured by an IQ150 portable pH-conductivity meter (Spectrum Technologies, Inc., Aurora, IL, USA).
The plant samples were washed three times using distilled water and oven-dried. The samples were ground, sieved through 100 mesh sieve, and digested with an H 2 O 2 -HClO 4 -HF-HNO 3 method in PTFE jars on an electric heating plate. Two Chinese national standard samples (GBW07427, GBW10048), as well as a blank sample was used as references to control the quality of the sample analysis. The total concentrations of cadmium (Cd), lead (Pb), copper (Cu), chromium (Cr), and zinc (Zn) in plants, soil and gangue were measured by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, IRIS Intrepid II XSP, Thermo Scientific, Waltham, MA, USA). The recovery rates of the standard samples were within 90-110%.

Potential Ecological Risk Assessment
The concept of potential ecological risk assessment (RI) was introduced by Hakanson in 1980, and is widely utilized in soil and sediment HM studies. The calculation formulas are as follows: RI is the sum of all indexes of five HMs, C i f is the pollution factor, C i D is the measured concentration, C i R is a reference value, which is set as Henan HM background content (Table 1), E i r is the monomial potential ecological risk index, and T i r is the toxic factor for HM which is 30, 5, 5, 2 and 1 for Cd, Pb, Cu, Cr and Zn. Five categories of monomial risk are low potential ecological risk (E i r < 40), moderate potential ecological risk (40 ≤ E i r < 80), considerable potential ecological risk (80 ≤ E i r < 160), high potential ecological risk (160 ≤ E i r < 320) and very high ecological risk (E i r ≥ 320). Four categories of RI value are low ecological risk (RI < 150), moderate ecological risk (150 ≤ RI < 300), considerable ecological risk (300 ≤ RI < 600) and very high ecological risk (RI ≥ 600) [26].

Screening of Phytoextractors
The prior objective of the remediation is to reduce the content of toxic HMs in soil. The following factors related to phytoremediation should be considered.

BCF Calculation
BCF indicates the HM uptake efficiency of a plant species into its tissues from the surrounding environment [27]. It is calculated as follows [28]: where C tissue is the concentration of the target HM in the plant harvested tissue and C soil is the concentration of the same HM in the soil or other substrate.

TF Calculation
TF indicates the efficiency of the plant in translocating the accumulated HM from its roots to shoots. It is calculated as follows [29]: TF = C AG /C root (5) where C AG is concentration of the HM in plant aboveground part and C root is concentration of the HM in plant roots.

Adaptability Factor
Adaptability factor indicates the dominance of the phytoextractor candidates at sites of different pollution levels. Two categories for pollution level are set-moderate risk site (MRS) for the soil RI under 300, considerable risk site (CRS) for the soil RI over 300. Three categories for dominance are set-rare (R, dominance < 50%), common (C, 50% < dominance < 75%), and dominant (D, dominance > 75%). Overall, four categories are set to evaluate the adaptability factor of the plants: (G1) the dominance of the plants was D at CRS sites; (G2) the dominance was C at CRS sites; (G3) stood for D dominance status at MRS sites; (G4) contained the rest of the plants, including the species of dominance status C at MRS sites and dominance status R.

Statistical Analyses
The mean values and standard deviation were calculated by using Microsoft Excel (Microsoft, Redmond, WA, USA). Statistical significance of differences among means was determined by one-way ANOVA using R language, taking p ≤ 0.05 as significant level and Tukey Contrasts multiple comparisons of means were carried out for pairwise analyses.

Soil HM Contamination
The concentration of the soil Cd, Pb and Zn exceeded the corresponding local soil HM background value ( As shown in Figure 2, the highest RI was up to 627.30. There were 1, 5, 5 and 2 samples for very high risk, considerable risk, moderate risk and low risk, respectively. Figure 2 also shows that the contribution of Cd monomial risks varied from 85.48 percent to 96.48 percent with an average of 90.51 percent among all the samples. The results of potential ecological risk assessment indicated the gangue dump and the adjacent land was contaminated by Cd, Cu, Pb and Zn, and Cd contributed the most of the RI. Cd exposure could result in carcinogenicity, mutagenicity, and teratogenicity, endocrine disruption, biological calcium regulation interference, renal failure and chronic anemia [31][32][33]. In addition, the reference dose (RfD) of oral soil digestion is provided by USEPA and Leung et al. [34,35], which is 0.001, 0.003, 0.04, 0.0035 and 0.3 mg·(d·kg −1 ) for Cd, Cr, Cu, Pb and Zn. The RfD of Cd is much lower than those of the others. Also, some researchers stated that Cd is the HM of most concern because it is the "only metal that might pose human or animal health risks at plant tissue concentrations that are not generally phytotoxic" [36]. Along with the potential ecological index assessment result, Cd should therefore be considered as the target element for the screening of native phytoextractors.

Statistical Analyses
The mean values and standard deviation were calculated by using Microsoft Excel (Microsoft, Redmond, WA, USA). Statistical significance of differences among means was determined by oneway ANOVA using R language, taking p ≤ 0.05 as significant level and Tukey Contrasts multiple comparisons of means were carried out for pairwise analyses.

Soil HM Contamination
The concentration of the soil Cd, Pb and Zn exceeded the corresponding local soil HM background value ( As shown in Figure 2, the highest RI was up to 627.30. There were 1, 5, 5 and 2 samples for very high risk, considerable risk, moderate risk and low risk, respectively. Figure 2 also shows that the contribution of Cd monomial risks varied from 85.48 percent to 96.48 percent with an average of 90.51 percent among all the samples. The results of potential ecological risk assessment indicated the gangue dump and the adjacent land was contaminated by Cd, Cu, Pb and Zn, and Cd contributed the most of the RI. Cd exposure could result in carcinogenicity, mutagenicity, and teratogenicity, endocrine disruption, biological calcium regulation interference, renal failure and chronic anemia [31][32][33]. In addition, the reference dose (RfD) of oral soil digestion is provided by USEPA and Leung et al. [34,35], which is 0.001, 0.003, 0.04, 0.0035 and 0.3 mg·(d·kg −1 ) for Cd, Cr, Cu, Pb and Zn. The RfD of Cd is much lower than those of the others. Also, some researchers stated that Cd is the HM of most concern because it is the "only metal that might pose human or animal health risks at plant tissue concentrations that are not generally phytotoxic" [36]. Along with the potential ecological index assessment result, Cd should therefore be considered as the target element for the screening of native phytoextractors.

Screening for Native Phytoextractors
BCF is a more important measure than shoot metal concentration when considering the potential of a given candidate species for phytoextraction [23]. As shown in Table 5, the plant samples with BCF greater than 1 were mainly for Cd element. The highest one was in AG sample of Artemisia roxburghiana Bess. (8.22), and the lowest one was in maize grain (0.27). Only a few samples had BCF higher than 1 for the other HM elements.     Accumulation and exclusion are two tolerance strategies evolved by plants to live in HM polluted medium [42][43][44][45]. HM accumulators are defined by the TF > 1, while root HM transport to shoot in excluders is restricted and TF < 1 [43,46,47]. A TF value greater than 1 indicates the translocation of HM from root to AG parts [48]. The translocation factors of the studied plant species were shown in Table 6. The Cd TFs of Artemisia roxburghiana Bess., Artemisia scoparia Waldst. Et Kit., Bidens pilosa Linn., and Artemisia argyi Lévl. et Van. were higher than 1. They could be considered as Cd accumulators. The other ones behaved as Cd excluders, suggesting Cd immobilization in the roots. The dominance of species is a combination of abundance and cover of plants, which indicates the biomass and adaptability in certain habitats. With the introduction of the habitat classification by the potential ecological risk assessment, the adaptability factor can reflect the plants' vitality in the environment under different levels of HM stress. The adaptability factor could be a useful supplementary criterion other than BCF and TF for phytoextractor screening from a community ecological point of view, though it is a "black box" of the plants' adaptability indication, and lacks quantitative tolerance study like the response of biomass, length, chlorophyll (CHL), superoxide dismutase (SOD), peroxidase (POD), malondialdehyde (MDA), and soluble protein (SP) within the plants under HM stress. Furthermore, DNA changes determined by random amplified polymorphic DNA (RAPD) could to be employed as a useful tool to identify the genotoxic effects caused by HMs within plant individuals [49]. Therefore, the results obtained from the field survey should be validated with the methods mentioned above.
The previous study stated that adverse effects were observed in plant tissues stressed by high HM content, such as reduction of biomass, length and total protein contents [50]. In this study, all plant species in Table 2  Herbs are more promising for phytoextraction than shrubs and trees due to their high growth rate, high biomass, and more adaptability under stress [51]. Therefore, the following herbs like Bidens pilosa Linn., Artemisia argyi Lévl. et Van., Artemisia scoparia Waldst. Et Kit., Artemisia roxburghiana Bess., and Arthraxon hispidus (Trin.) Makino were chosen to be discussed based on bioconcentration factors, translocation factors and adaptability factors analyzed above.
Bidens pilosa Linn. belonged to G1. The Cd content, BCF and TF was 1.16 ± 0.30 mg·kg −1 , 1.41, and 3.05, when the root soil concentration was 0.82 ± 0.57 mg·kg −1 . According to Yoon et al. [10], plant species with both BCF and TF greater than 1 have the potential to be used for phytoextraction. Bidens pilosa met this criterion in this study. Bidens pilosa was considered as a hyperaccumulator under laboratory conditions in previous studies [52]. The Cd concentration of AG was higher than 100 mg·kg −1 , the BF and TF values were all greater than 1.0, when soil Cd concentration was 8 and 16 mg·kg −1 in the form of CdCl 2 solution. The roots were able to uptake HM from aqueous phase. However, soil HMs in the field were mainly insoluble and unavailable for uptake by plants due to strong binding of HM ions to soil particles and precipitation [4]. This could explain the BCF difference between present study and the laboratory experiment. In addition, the biomass showed no significant reduction and the HM stress did not show any significant impact on CHL, SOD, POD, MDA, and SP, which showed high tolerance to Cd [52]. The high adaptability at polluted sites, BCF, TF greater than 1, and the potential to be a hyperaccumulator suggested Bidens pilosa Linn. could be used as a native phytoextractor for soil Cd phytoremediation.
In the previous study [53] The results showed it could be considered as a potential phytoextractor. The difference of accumulation performance could be explained by the lower root soil Cd content as well as the lower bioavailability caused by the difference between the natural soil and the soil-HM compound formulated in the laboratory.
Artemisia scoparia Waldst. Et Kit. was reported having certain ability to extract HM from contaminated soil in previous research [54]. Its AG content, BCF and TF of Cd were 3.33 ± 0.55 mg·kg −1 , 0.69 and 1.25 in the soil with Cd concentration of 4.76 ± 1.2 mg·kg −1 at a mine-affect area, which did not meet the requirement for phytoextractor due to the low BCF. According to another investigation [55], Artemisia scoparia Waldst. Et Kit. from different sites showed quite contradictory phytoextraction performances. The AG content, BCF and TF of the sample from site 3 were 4.02 ± 0.89 mg·kg −1 , 10.05 and 4.14, while the parameter from site 4 were 0.34 ± 0.12 mg·kg −1 , 1.31 and 0.79. The total soil Cd concentration were similar at two sites, which were 0.30 ± 0.04 for site 3 and 0.26 ± 0.05 mg·kg −1 for site 4. The difference between DTPA extractable Cd (0.04 ± 0.004 for site 3 and 0.01 ± 0.001 mg·kg −1 for site 4) suggested that the phytoextraction characteristics of Artemisia scoparia Waldst. Et Kit. relied on the DTPA extractable Cd in the soil and might lack the ability to mobilize carbonate-bound and residual state Cd. Artemisia scoparia Waldst. Et Kit. was divided into G1 group in present study. Its AG content, BCF and TF were 2.62 ± 0.16 mg·kg −1 , 5.82 and 5.70, which had the potential as a phytoextractor on the sites with high bioavailable Cd content.
Artemisia roxburghiana Bess. belonged to G3 in this case. The soil pollution level was moderate: 23.96 ± 4.81, 12.25 ± 3.29, 52.55 ± 15.37 and 0.18 ± 0.06 mg·kg −1 for Pb, Cu, Zn and Cd. The BCF and TF of AG for Pb, Cu, Zn and Cd was 0.27 and 4.11, 1.57 and 2.85, 1.67 and 4.48, 8.22 and 6.73, respectively. The lack of dominance at CRS sites might limit its remediation application. Lei and Duan [56] found that Artemisia roxburghiana Bess.'s BCF for Pb, Zn and Cu were less than 0.005, and the HM content remained a low level at very highly polluted sites (over 7000, 35,000 and 400 mg·kg −1 for Pb, Zn and Cu), showing weak phytoremediation capacity. However, Artemisia roxburghiana Bess. was also found to have the ability to reduce a considerable amount of Pb, Zn and Cu concentration and increase the soil fertility in rhizosphere simultaneously. Although Artemisia roxburghiana Bess. was unfit for phytoextraction at considerable risk sites, it showed the potential to treat the soil at moderate risk sites due to the high BCF and TF as well as the ability of soil fertility improvement.
Arthraxon hispidus (Trin.) Makino was classified as one of G1 species in this study. The Cd content, BCF and TF of AG was 0.54 ± 0.05 mg·kg −1 , 0.59 and 0.39. Our results were in consistency with the previous experiment conducted in Cd(NO 3 ) 2 treated soil with 3.00 mg·kg −1 of total Cd. The BCF and TF of AG was 4.17 and 0.95, also showing the exclusion characteristics [57]. The BCF difference between the two study could be explained by high bioavailable Cd content due to the implementation of Cd(NO 3 ) 2 during the sample formulation. The exclusion strategic of Arthraxon hispidus (Trin.) Makino suggested it had little potential for phytoextraction, but it had the potential to be used for phytostabilization at considerable risk sites.

Conclusions
The concentration of Cd, Pb, Cu and Zn at the study site exceeded the local background value and corresponding CNS, The RI values were over 300 in 46.2% of the sediment samples, the maximum RI was 627.80, which indicated very high ecological risk. Cd contributed the most to the RI, and was the main contaminant as well as the prior target for soil remediation. There were 24 familiae, 37 genera and 40 species of plants in the study area, and the community structure is simple. Bidens pilosa Linn. and Artemisia argyi Lévl. et Van. had high potential to be phytoextractors due to their high AG Cd concentration, BCF, TF and dominance at considerable risk sites. Artemisia roxburghiana Bess. was fit for Cd phytoextraction at moderately risk sites. Arthraxon hispidus (Trin.) Makino was classified as a Cd excluder, which seemed suitable for phytostabilization. Further studies should be conducted under controlled environment, such as the responses of plants' BCF, TF, biomass, CHL, SOD, POD, MDA, SP and DNA changes, as well as their influence on the soil organic matter, pH, and HM bioavailability.