Introduction

Heavy metals are environmental contaminants that are hazardous to humans and other biota (Marques et al. 2009). Heavy metal contamination of soils has become a serious environmental problem in industrial and agricultural areas (Bhargava et al. 2012). Mining is one of the most important sources of heavy metal contamination (Li et al. 2014). For example, overburden and waste materials from mining, with their consequent environmental risks, are contaminants in surrounding areas (Acosta et al. 2011; Parra et al. 2014). Furthermore, mine tailings are difficult to revegetate because of high concentrations of metals, low nutrient levels, abnormal pH values, and low water retention capacity (Cano-Reséndiz et al. 2011; Solís-Dominguez et al. 2012). In most cases, the tailings also have steep slopes (Cano-Reséndiz et al. 2011). In China, mining activities have generated about 1.5 million ha of wasteland (Zhuang et al. 2009), and the area is increasing at 46,700 ha annually (Li et al. 2007). Thus, there is an urgent demand for effective management of mine tailings.

There is a variety of techniques for in situ reclamation of mine wastes, many with high maintenance costs and secondary pollution risk (Ucisik and Trapp 2008). Phytoremediation is a green technology that uses plants to mitigate environmental problems without the need for excavation and disposal of the contaminating material (Zhang et al. 2014). However, most tailing disposal sites are devoid of vegetation cover, principally as a result of the elevated concentration and bioavailability of heavy metals combined with adverse growth conditions (Mendez and Maier 2008a; Monterroso et al. 2014). Therefore, a crucial consideration in phytoremediation technology is the selection of plant species to meet the objectives of phytoremediation (Solís-Dominguez et al. 2012; Suchkova et al. 2014; Zhang et al. 2014). The ideal plant for phytoremediation should grow rapidly, produce high biomass, develop a deep root system, grow on poor soils, and show metal stress tolerance (Pérez-Esteban et al. 2013; Pottier et al. 2015). Relatively few plants possess all these attributes. However, woody species constitute the majority of plant biomass in native forests and shrublands, which can provide extensive canopy cover and establish a deep root network to prevent erosion of mine wastes in the long term (Mendez and Maier 2008a). Furthermore, trees can decrease metal mobility and toxicity through root growth (Brunner et al. 2008). Thus, research on suitable woody plants, and understanding their metal accumulation patterns, is critically needed to assess their potential use in phytoremediation in the field.

Few studies have evaluated the utility of woody species for phytoremediation of mine tailings or contaminated soil (Mertens et al. 2004; Domínguez et al. 2008; Seo et al. 2008; de Souza et al. 2012; Pottier et al. 2015). In recent years, willow species (Salix L. spp.), which show rapid growth and produce extensive root systems, have been suggested as suitable for use in the remediation of metal-contaminated soils. Willows typically accumulate the highest contents of heavy metal in wood compared with other tree species (Algreen et al. 2014; Evlard et al. 2014). A number of willow species have been studied for their ability to tolerate and accumulate heavy metals (Zacchini et al. 2009, 2011; Pietrini et al. 2010; Evlard et al. 2014; Wang et al. 2014). The genus Quercus L. (oak species) is a keystone taxon from both ecological and economic perspectives (Aldrich and Cavender-Bares 2011). Compared with willow species, knowledge of the physiological responses of oak trees to heavy metal stress is incomplete (Gogorcena et al. 2011). Oak trees have been reported as more tolerant than willow trees to growth in metal-contaminated soils (Evangelou et al. 2012). Moreover, oak trees have proven suitable for reforestation in areas contaminated with heavy metals (Prasad and Freitas 2000; Domínguez et al. 2009; Zhao et al. 2015). The red oak species Shumard oak (Q. shumardii Buckley) and willow oak (Q. phellos L.), and the white oak species southern live oak (Q. virginiana Mill.), from southern USA, were introduced to China for the establishment of plantations in different sites and soil types in the Yangzi River Delta in 2002. These oak species show good adaptability to natural conditions in the delta, and trees began to bear fruit at age 6–9 years. S. matsudana Koidz. and S. integra Thunb. are native willow species in China and are widely planted throughout China. These five tree species were tested as candidates for phytoremediation by growing seedlings in lead/zinc (Pb/Zn) mine tailings. The growth, tolerance index, root traits, chlorophyll parameters, and metal uptake of seedlings were assessed. The objectives of this study were to elucidate the adaptive capabilities of the selected oak and willow species for growth on mine tailings and their potential use for phytoremediation of mine tailings in China.

Materials and methods

Site description and soil samples

Tailings from a Pb/Zn mine were collected from the city of Hangzhou (30° 126′ N, 119° 847′ E), China. The control medium (river sand), which was unaffected by heavy metal contamination or mining activities, was collected from Hangzhou (30° 057′ N, 119° 956′ E). River sand was chosen because its texture is similar to that of the mine tailings. All areas where materials were collected for this study experience a subtropical climate. The physicochemical properties of samples from the uppermost 30 cm of the mine tailings and river sand were analyzed (Table 1). According to the Chinese Environmental Quality Standard for Soils (GB 15618-1995), the Class 3 standards of maximum concentrations of Pb, Zn, and cadmium (Cd) permitted in soils are 500, 500, and 1.0 mg kg−1, respectively. However, heavy metal concentrations in the mine tailings were much higher than the recommended limits.

Table 1 Characteristics of three types of medium
Full size table

Plant materials and growth conditions

Seeds of Q. shumardii, Q. phellos, and Q. virginiana were collected from plantations in Shanghai, China. The seeds were sown in a pot made of non-woven fabric (diameter 6 cm × height 10 cm) that contained growing nutrient medium comprising perlite/peat (1:3; v/v). When they reached the stage of eight to ten leaves, uniform seedlings were selected and transplanted into black cylindrical plastic pots (diameter 18 cm × height 23 cm) without fertilizer for the experiment. Stem cuttings of S. integra and S. matsudana were obtained from native habitats in Shandong and Zhejiang provinces. Healthy willow stem cuttings of uniform length (8 cm) were rooted in pots made of non-woven fabric (diameter 6 cm × height 10 cm). After growth for 2 months, uniform rooted cuttings of similar height and weight were selected for the experiment.

Experimental method

The experiment was performed in the greenhouse of the Research Institute of Subtropical Forestry, Chinese Academy of Forestry, Hangzhou, China. The river sand or tailings were air-dried, blended, sieved, and then 3-kg samples were placed in the black cylindrical plastic pots without fertilizer (diameter 18 cm × height 23 cm). Uniform seedlings/rooted cuttings (S. integra, S. matsudana, Q. shumardii, Q. phellos, and Q. virginiana) with their nutrient medium were transplanted into the pots in April 2011. Each pot contained 2 seedlings or rooted cuttings. Each species was grown in control medium, 50 % tailings substrate, and 100 % tailing substrate (river sand/tailings at 1:0, 1:1, and 0:1, v/v, respectively), resulting in 15 treatments each comprising 5 pots. Each of the 15 treatments was an experimental unit, and the 3 experimental units were arranged in a randomized block design. The plants were cultivated for 150 days with natural light, day/night conditions of 25–35/20–25 °C, and relative humidity of 70/85 %. During the growth period, irrigation was scheduled according to tensiometer measurements to maintain the soil moisture content at approximately field capacity. Metal-related phytotoxicity symptoms were recorded during the experiment. All seedlings were harvested after 150 days.

Biomass measurements

After harvesting, the leaves, stems, cuttings, roots in nutrient medium (inner roots), and roots in river sand or tailings (external roots) were excised and separated. The samples were washed thoroughly with distilled water. The roots were immersed in 20 mmol L−1 Na2-EDTA for 15 min to remove metals adhering to the root surface (Wang et al. 2014). The dry weights of the samples were measured after drying at 75 °C for 3 days to a constant weight. To compensate for the marked differences in initial plant development between species, relative growth rate (RGR) was adopted for calculation of total plant dry weight. The RGR from initiation of the experiment to final harvest was calculated in accordance with the formula of Dimitriou et al. (2006):

$$ RGR=\frac{\left( \ln {A}_{\mathrm{f}}- \ln {A}_{\mathrm{i}}\right)}{t_{\mathrm{f}}-{t}_{\mathrm{i}}} $$
(1)

where A f is measured plant dry biomass at the final harvest calculated as the mean value of three replicates per species (n = 3), A i is measured plant dry biomass at the initial harvest calculated as the mean of six plants’ dry biomass per species (n = 6), and t is time in weeks.

To assess the tolerance of the six species to mine tailings, the tolerance index (TI) based on biomass was calculated for each treatment as TI = B t/B c (Metwally et al. 2005), where B t (g plant−1) is treatment biomass and B c (g plant−1) is control biomass. High values indicate high tolerance by plants.

Measurement of root traits

All roots from an individual plant were scanned while fresh using a root positioning system/STD4800 scanner (Regent Instruments Inc., Québec, Canada) at the end of the experiment. From the images obtained, the characteristics of roots (mean total length, mean surface area, mean volume, mean diameter, and mean number of root tips) were analyzed using WinRHIZO Pro 2005b software (Regent Instruments).

Estimation of chlorophyll content

We used an Opti-Sciences CCM-200 chlorophyll content meter (Opti-Sciences Inc., Hudson, NH, USA) to non-destructively estimate the chlorophyll content in leaves by recording the chlorophyll concentration index (CCI).

Estimation of chlorophyll fluorescence parameters

Chlorophyll fluorescence parameters for mature leaves (third–fourth leaf from the apex) from plants of the control and each treatment was measured after 150 days, with measurement performed during 8:30–11:00 a.m. In vivo chlorophyll fluorescence was measured on the upper surface of mature leaves at room temperature after dark adaptation for 25 min using a portable fluorometer (Mini-PAM, Walz, Effeltrich, Germany). The minimal fluorescence level in the dark-adapted state (F 0) was measured when the modulated light was sufficiently low (<0.1 μmol m−2 s−1) to avoid inducing significant variable fluorescence. The maximal fluorescence level in the dark-adapted (F m) and light-adapted (F m′) states was determined before or after addition of the actinic light by 0.8 s of saturating white light (160 μmol m−2 s−1) to close all reaction centers and drive photochemical quenching to zero. The steady-state value of fluorescence (F s) under actinic light was also recorded. Using both light- and dark-adapted fluorescence parameters, we calculated the maximum efficiency of photosystem II (PSII) photochemistry (F v/fm) in the dark-adapted state, the quantum yield of PSII electron transport as ΦPSII = (F m′ – F s)/fm ′, photochemical quenching as qP = (F m − F s)/(F m′ − F 0), and non-photochemical quenching as NPQ = (F m′ − F m)/fm ′ (Kumar and Prasad 2015).

Chemical analysis

Dried plant samples (shoots and roots) were ground into powder, and 0.2 g of each sample was digested with a mixture of 4 mL of nitric acid (HNO3) and 1 mL of perchloric acid (HClO4). Metal concentrations (Cd, Pb, and Zn) were determined using atomic absorption spectroscopy (AAS) (SOLAAR M6, Thermo Fisher Scientific Inc., Waltham, MA, USA). Tailings and soil samples were air-dried and then sieved through a mesh (<2 mm) to yield a homogeneous mixture. Total Pb, Zn, and Cd concentrations were determined using flame AAS after extraction with 5 mL of a 4:1 (v/v) mixture of 65 % HNO3 and 70 % HClO4.

River sand and tailing samples were air-dried and sieved (<2 mm) after sampling from the mine tailing areas. Samples were then analyzed for total and available metals (Pb, Zn, and Cd) and hydrolysable nitrogen, available phosphorus, and rapidly available potassium content. Hydrolyzable nitrogen was determined by the Berthelot reaction method and available phosphorus by molybdenum blue method after digestion with 1.0 g of potassium sulfate and 5 mL of concentrated sulfuric acid (Deng et al. 2004). Total Pb, Zn, Cu, and K contents were determined using AAS after extraction with 5 mL of a 4:1 mixture (v/v) of 65 % HNO3 and 70 % HClO4 (Shu et al. 2001). The pH values (solid/distilled water of 1:2.5; v/v) of river sand/tailing samples were measured. Organic matter was determined according to Paniz et al. (2001). Certified reference materials (mixed shoots of shrubs from Pb/Zn mine tailings; GBW 07602, China) were used to ensure the quality of analyses. Analyses of the element concentrations in the plant material were performed at the Quality Testing Center for Edible Forest Products of State Forestry Administration (Hangzhou, China).

The bioconcentration factor (BCF) was used to estimate a plant’s ability to accumulate metals from soils. BCF was defined as the ratio of metal concentration in plant roots or shoots to that in soil. BCF at the end of the experiment was calculated as BCF = A tissues/A soil (Deng et al. 2004), where A tissues (mg kg−1) is the total heavy metal accumulated in roots or shoots, and A soil (mg kg−1) is the heavy metal concentration in the soil.

The translocation factor (TF) can be used to estimate a plant’s ability to translocate metals from the roots to the shoots. The TF at the end of the experiment was calculated as TF = A s/A r (Deng et al. 2004), where A s and A r are total heavy metals accumulated in shoots and in roots, respectively (both in mg kg−1).

Statistical analysis

The data are expressed as the mean of at least three replications ± standard error. All statistical tests were performed with SPSS 19.0 (SPSS Inc., Chicago, IL, USA). The fitting of the data to a normal distribution for all parameters determined was checked with the Kolmogorov–Smirnov test. When necessary, analytical data were transformed using logarithms to assure normal distribution (Parra et al. 2014). For all experimental variables, two-way analysis of variance was applied with tailings and species as factors and the least significant difference test was used for comparisons of means. Differences were considered significant when the analysis of variance F-test achieved p < 0.05.

Results

Biomass

The biomass of plants grown in each treatment is shown in Table 2. The growth of all species, except Q. virginiana, was inhibited by the tailings. Compared with the control, the biomass of Q. virginiana grown in 100 % tailings increased slightly, but the biomass in different treatments showed no differences. Q. shumardii produced the highest leaf, root, and total biomass among all species. However, compared with the control, the biomass of leaves, stems, and external roots in 100 % tailings decreased significantly (p < 0.05). The performance of S. matsudana was similar to that of Q. shumardii; the biomass of S. matsudana was reduced by 36.6 % in 100 % tailings, which was a greater reduction than for the other species. Relative to controls, the biomass of Q. phellos and S. integra grown in 50 and 100 % tailings was reduced slightly, but the biomass of Q. phellos did not significantly differ for the different treatments. RGR showed the same change trends as described above for dry biomass.

Table 2 Dry weight (DW) (g) of root and aboveground tissues of five woody species grown in mine tailings for 150 days
Full size table

Tolerance of the five species in tailings were assessed by calculating TI values (Fig. 1), which differed significantly between treatments (p < 0.05). In general, the TI values of the five species were similar in 50 % tailings. However, the low TI value of S. matsudana indicated greater sensitivity than did the other species in 100 % tailings (p < 0.05), and TI of 0.97–1.04 of Q. virginiana was the most tolerant of the tested species.

Fig. 1
figure 1

Tolerance index (TI) of five woody species grown in Pb/Zn mine tailings. Each value represents the mean of three (n = 3) replicates ± SE. Different letters on the bars for the same species indicate significant difference between the treatments (p < 0.05). TI values of five species, species ****; tailings ****; species × tailings ***. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001

Full size image

Root traits

After exposure to heavy metals, the roots of all plants exhibited different levels of blackening, and the symptoms became more severe with increasing heavy metal concentrations. The root systems of Quercus spp. and Salix spp. showed marked differences among all treatments because of the different propagation materials. In general, Quercus spp. had well-developed taproots but variable lateral root development. For example, roots of Q. shumardii developed throughout the whole container, whereas roots of Q. virginiana and Q. phellos only partially filled the container (Table 3). The root systems of the two Salix spp. filled the whole container owing to the development of adventitious roots in control and 50 % tailings treatment. The root characteristics of all species grown in tailings varied significantly (Tables 2 and 3, p < 0.05) depending on species and tailings concentration. The root biomass of Salix spp. and Q. shumardii grown in 100 % tailings decreased significantly compared with controls. In addition, the external root morphological parameters of Salix spp. and Q. shumardii were reduced significantly by 100 % tailings (except for average root diameter). In general, reductions in root biomass and morphological parameters were more pronounced in external than inner roots.

Table 3 Root characteristics of five woody species grown in mine tailings for 150 days
Full size table

Chlorophyll concentration index

After 150 days of heavy metal exposure, foliage of S. matsudana and Q. shumardii treated with 100 % tailings had begun to yellow, whereas leaves of other species showed no pronounced toxicity symptoms. The CCI values at harvest were decreased by tailings treatment for all species, and there were significant differences among all species (p < 0.05; Fig. 2).

Fig. 2
figure 2

Chlorophyll concentration index (CCI) of leaves (third to fourth leaf from the apex) of 5 wood species grown in mine tailings for 150 days. Each value represents the mean of 15 (n = 15) replicates ± SE. Different letters on the bars for the same species indicate significant difference between the treatments (p < 0.05). CCI values of five species, species ****; tailings *; species × tailings ***. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001

Full size image

Chlorophyll fluorescence

The leaf F 0 value after 150 days was lower in tailing-treated Salix spp. compared with controls (Fig. 3). However, F 0 for Quercus spp. showed an opposite trend. Furthermore, the tailing treatment did not significantly affect the F m value (Fig. 3). F v/fm reflects the potential quantum efficiency of PSII. The F v/fm value of Q. shumardii showed a linear decrease with tailing treatment (p < 0.05; Fig. 3), with significant decreases in F v/fm under treatment with 50 % (1.5 %) and 100 % tailings (2.8 %). However, F v/fm for the other species showed no significant difference in all treatments. The ΦPSII values for S. integra and Q. phellos did not significantly differ among treatments. In contrast, ΦPSII values of S. matsudana, Q. shumardii, and Q. virginiana were significantly lower in tailing treatments compared with controls. qP gives an indication of the proportion of PSII reaction centers that are open, and the trend for qP (Fig. 3) was very similar to that of ΦPSII. In contrast, for 100 % tailings, NPQ values changed slightly in S. matsudana and Q. virginiana and increased notably in all other species; only Q. phellos also increased for 50 % tailings (Fig. 3). In addition, S. matsudana had significantly higher NPQ values than did the other species.

Fig. 3
figure 3

Chlorophyll fluorescence parameters in the leaves (third to fourth leaf from the apex) of five species grown in mine tailings for 150 days. Each value represents the mean of nine (n = 9) biological replicates ± SE. Different letters on the bars for the same species indicate significant difference between the treatments (p < 0.05). F 0 of five species, species ****; tailings ns; species × tailings **. Fm of five species, species ****; tailings ns; species × tailings ns. Fv/Fm of five species, species ****; tailings ns; species × tailings ns. ΦPSII of five species, species ****; tailings ****; species × tailings ***. NPQ of five species, species ****; tailings ****; species × tailings ***. qP of five species, species ****; tailings ****; species × tailings ****. ns not significant; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001

Full size image

Accumulation and translocation of heavy metals in plants

The concentrations of Pb, Zn, and Cd in tissues of plants grown in tailings are shown in Fig. 4. Generally, tissues of all plant species grown in tailings contained higher metal concentrations than did those grown in river sand (p < 0.05). There were also higher metal concentrations in roots than in shoots for all species grown in the same tailing treatment, except for Cd and Zn concentrations in S. integra (p < 0.05). The Cd concentrations in shoots of Quercus spp. were below the limit of detection. Plants grown in tailings contained higher Pb and Zn concentrations in their tissues than Cd. Furthermore, the Pb and Zn concentrations in roots were higher for 100 than 50 % tailings, except for Q. shumardii. However, Pb and Zn concentrations in S. integra shoots showed an opposite trend to that for roots. The Cd concentrations in tissues were higher in 50 than 100 % tailings, except for Q. virginiana (Fig. 4).

Fig. 4
figure 4

Average Pb, Zn, and Cd concentrations (mg kg−1) in dry plant tissues of five species exposed to mine tailings for 150 days. Each value represents the mean of three (n = 3) replicates ± SE. Different letters on the bars for the same species of roots and shoots indicate significant difference between the treatments respectively (p < 0.05). Pb concentration of shoots, species ****; tailings ****; species × tailings ****. Pb concentration of roots, species ****; tailings ****; species × tailings ****. Zn concentration of shoots, species ****; tailings ****; species × tailings ****. Zn concentration of roots, species ****; tailings ****; species × tailings ****. Cd concentration of shoots, species ****; tailings ****; species × tailings ****. Cd concentration of roots, species ****; tailings ****; species × tailings ****. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001

Full size image

The BCF and TF were used to further evaluate the ability of the five species to accumulate heavy metals (Fig. 5). In general, the BCF and TF values of the five species were <1, except for S. integra grown in tailings, which had the highest TF values for Zn (TF = 1.42–2.18) and Cd (TF = 1.03–1.45). For all metals, BCF values of roots were higher for 50 than 100 % tailings in all species, except for Q. virginiana and the BCF value of Zn for Q. phellos. BCF of shoots showed a similar trend to BCF of roots, except for Pb in Q. shumardii; however, BCF for Cd could only be calculated for Salix spp. The BCF values of Zn were higher than those of Pb (p < 0.05). Finally, TF for Zn showed a trend similar to BCF, with lower values in 100 % tailings. The TF values ranged around 0.5 with the exception of TF >1 for S. integra. In contrast, TF values were generally lower for Pb than Zn and decreased with increasing proportion of tailings for S. integra and Q. virginiana but increased for S. matsudana and Q. shumardii.

Fig. 5
figure 5

Translocation factor (TF) and bioconcentration factors (BCF) of Pb, Zn, and Cd in five woody species. Each value represents the mean of three (n = 3) replicates ± SE. Different letters on the bars for the same species indicate significant difference between the treatments (p < 0.05). TF values of Pb, species ****; tailings ***; species × tailings ****. TF values of Zn, species ****; tailings ****; species × tailings ****. TF values of Cd, species ****; tailings ****; species × tailings ****. Shoot BCF values of Pb, species ****; tailings ****; species × tailings ****. Shoot BCF values of Zn, species ****; tailings ***; species × tailings ****. Shoot BCF values of Cd, species ****; tailings ***; species × tailings ****. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001; Root BCF values of Pb, species ***; tailings ***; species × tailings ***. Root BCF values of Zn, species ***; tailings ***; species × tailings ***. Root BCF values of Cd, species ****; tailings ***; species × tailings ****. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001

Full size image

Discussion

Heavy metal tolerance of woody species

Previous studies confirmed that plant populations surviving in areas contaminated by heavy metals are differentiated from other populations of the same species by a genetically based tolerance (Haliru et al. 2009). The Quercus spp. and Salix spp. in this experiment were able to survive in mine tailings. However, the five species responded differently to high concentrations of heavy metals in tailings. The different abilities to maintain normal growth in tailings reflect differences in resistance or tolerance of plants to metal toxicity (Wang et al. 2014). Growth inhibition resulting in reduction in biomass production is common in plants exposed to high metal concentrations (de Souza et al. 2012). In the present study, Q. virginiana grew well in tailings, with no apparent effect on measured plant growth parameters. The other four species showed reductions in biomass production when grown in tailings compared with river sand. These findings indicate that Q. virginiana had relatively high tolerance to heavy metals compared with the other four species, especially when grown in 100 % tailings. After exposure to tailings, particularly 100 %, S. matsudana and Q. shumardii plants were stunted compared with controls. After approximately 75 days of exposure to tailings, the foliage of these two species began to yellow and wilt from the leaf tip, with more severe symptoms at harvest (data not shown). The Q. phellos and S. integra showed no leaf wilting or sign of dehydration. The growth measures used in the present study indicated that Q. virginiana was better adapted to metal-contaminated soils.

Seedlings of Quercus spp. and rooted cuttings of Salix spp. showed relatively high tolerance to Pb, Zn, and Cd. Some chlorophyll fluorescence parameters (F m and F v/fm) remained unchanged in all species (except for decreases in Q. shumardii), with values of F v/fm slightly lower than the optimum (0.83) measured for most plant species (Maxwell and Johnson 2000). This suggested that heavy metal accumulation in leaves caused a reduction but not complete failure of photosynthesis (He et al. 2011). In the present study, the decrease in F v/F m value of Q. shumardii can be explained by the negative effect of tailings on efficiency of the PSII photochemical reaction and electron transport chain (Kumar and Prasad 2015). The results also indicated that reoxidation of the primary electron acceptor was limited, which results in a decrease in electron transport from PSII to PSI (Mallick and Mohn 2003; Kumar and Prasad 2015). ΦPSII was similar in all substrates only for S. integra and Q. phellos and decreased in all other cases for substrates with tailings. The qP value decreased particularly in leaves of plants grown in 100 % tailings, indicating that the number of closed or inactive PSII reaction centers increased, with lower use of light in photochemical reactions and a higher NPQ. This response signifies that light absorption exceeded the capacity of photosynthetic electron transport rate and carbon dioxide fixation (Vassilev and Manolov 1999) with the use of NADP decreased as a result of heavy metal stress (Vernay et al. 2008), resulting in the decline in plant biomass production. Similar trends were observed in maize (Vernay et al. 2008) and Talinum triangulare (Kumar and Prasad 2015). However, among Quercus spp., Q. virginiana demonstrated that the best response to all treatments and Q. shumardii was the most sensitive. Both Salix spp. performed well, although in 100 % tailings, S. integra maintained ΦPSII and partially also qP, increasing the energy dissipation capacity (i.e., NPQ).

Root growth was used as a measure of heavy metal tolerance because roots are the first organs to contact metals and therefore are likely the first to display symptoms of metal toxicity (Ma et al. 2014; He et al. 2015). In general, plant roots rapidly respond to absorbed heavy metals through reductions in root growth and changes in branching pattern (Elobeid et al. 2012; Wang et al. 2014). In this study, the external roots exhibited different levels of blackening, and growth was stunted after exposure to heavy metals, regardless of heavy metal concentration. The root system exhibits a certain phenotypic plasticity: plants with plastic root systems can often cope with a wide range of soil factors (Keller et al. 2003). Avoidance of zones contaminated by heavy metals is an efficient survival strategy, especially for trees (Dickinson et al. 1991). For example, in this study, the root systems of Q. phellos and Q. virginiana only partially filled the containers and no toxicity symptoms were observed. This indicated that adaptation of the root system morphology was also a mechanism for plants to tolerate contaminated soils. Additionally, the ability of the root system to pump heavy metals and to fit and expand into the contaminated zone are key factors for phytoextraction efficiency (Keller et al. 2003). A similar phenomenon was observed in the present study, in which the root systems of S. integra filled the whole container and accumulated relatively high concentrations of Zn and Cd in the root and aboveground tissues.

Metal accumulation and translocation

Metal accumulation in roots and shoots of the five studied species was much lower compared with other phytoaccumulators (Lu et al. 2014; Leal-Alvarado et al. 2016). However, concentrations of heavy metals in plant tissues were similar (Seo et al. 2008) or higher than those of other species (Evangelou et al. 2012; de Souza et al. 2012). Heavy metal concentrations were higher in roots than in shoots of plants grown in different media (except for Zn and Cd in S. integra). This is the usual behavior of non-hyperaccumulator species (Evangelou et al. 2012). For example, the Pb TF values of the five species were similar to that of Morus alba (Zhou et al. 2015) and lower than those of Erythrina speciosa (TF = 0.37–0.56) and Schizolobium parahyba (TF = 0.27–0.47) (de Souza et al. 2012). However, plant capacity for heavy metal translocation and accumulation is highly variable, depending on genotypic and environmental traits (Pietrini et al. 2010; Baccio et al. 2014). In Quercus spp. and S. matsudana, the highest absolute concentrations of Pb, Zn, and Cd were in roots, in 100 rather than 50 % tailings (except for Cd; Fig. 4). Nevertheless, S. integra showed higher Zn and Cd concentrations in shoots, especially when grown on 50 % tailings. However, the indexes of absorption and translocation indicated that the different species, with the exception of Q. virginiana, limited metal absorption when the concentration in the substrate was too high, as shown by the higher BCF values of roots in 50 % tailings. Translocation of Pb to the shoot was particularly limited in S. matsudana. Cd was not detectable in shoots of Quercus spp. and, in contrast, Zn was well translocated from soil to shoot since it is an important micronutrient. Among the species, S. integra showed very good translocation of all three metals to the aboveground part, especially in 50 % tailings. In addition, the TF index confirmed this result, demonstrating that S. integra may be a candidate for phytoextraction of Zn and Cd (Vysloužilová et al. 2003; Vassilev et al. 2005; Yang et al. 2015).

Excessive amounts of Pb, Zn, and Cd in soil may retard plant growth. The threshold for heavy metal toxicity is highly variable even among within plant species (Shi et al. 2015). Kabata-Pendias and Pendias (2001) provided some general values for phytotoxicity. In the cases of Pb, Zn, and Cd, normal and phytotoxic concentrations were 0.5–10 and 30–300 mg kg−1, 0.05–0.2 and 5–30 mg kg−1, and 10–150 and >100 mg kg−1 dry weight in shoot tissues, respectively. In the present study, most Pb, Zn, and Cd were retained in roots (except for Zn and Cd in S. integra), with concentrations approaching or higher than the normal or phytotoxic levels in roots of seedlings grown in Pb/Zn mine tailings. These findings suggested that the roots may have functioned as a barrier to heavy metal transport to shoots, which contributes to heavy metal tolerance in woody plants (de Souza et al. 2012). Root tissues accumulated significantly higher concentrations of metals than did shoot tissues (Fig. 4), indicative of high plant availability of the metals as well as their limited mobility once inside the plant (de Souza et al. 2012; Wang et al. 2014). This is an important tolerance mechanism that can be exploited for phytostabilization processes (Rabeda et al. 2015).

Implications for practice

According to Mendez and Maier (2008b), plants that show TF and BCF values of <1.0 and shoot concentrations of Pb, Zn, and Cd of <100, <500, and <10 mg kg−1, respectively, are appropriate for phytostabilization programs. For growth in tailings in temperate regions, plants are required to adapt to slightly anaerobic and wetland conditions (Craw et al. 2007). The five species of the present study are adaptable to wetland conditions. In arid and semiarid regions, drought-, metal-, and salt-tolerant plants are required (Mendez and Maier 2008a). Salix spp., and especially Q. virginiana, are salt-tolerant plants. Comparison of the TI values of the species in the present study and concentrations of heavy metals in shoots indicated that Q. virginiana was the most suitable for phytostabilization programs. From a toxicology perspective, this would be desirable to preclude metals from entering the food chain and so avoid potential environmental risks (Deng et al. 2004). According to Swiss regulations (FOEN 2005), the maximum allowed trace element concentrations in fodder (dry weight) are 150, 40, and 1 mg kg−1 for Zn, Pb, and Cd, respectively. The corresponding concentrations in shoots of Quercus spp. never exceeded these values (except for Zn of Q. phellos) and, because of high tannin concentrations in leaves, they are not eaten by deer and other animals (Evangelou et al. 2012), and thus pose a low risk for herbivores. S. integra accumulated relatively high concentrations of Zn and Cd in the aboveground tissues in our pot experiment. Through short rotations, S. integra could fulfill the purpose of phytoextraction from tailings in temperate regions driven by its high biomass production. However, the Zn and Cd concentrations in willow shoots surpassed the guideline limits, and the high Cd concentration represents a high risk for herbivores. S. matsudana and Q. shumardii displayed toxicity symptoms of leaf wilting and chlorosis when grown in 100 % tailings, and their sensitivity to heavy metal stress means they are unsuitable for phytoremediation under the experimental conditions.

Conclusions

The five tested tree species exhibited distinctly different responses of biomass yield when grown in tailings. However, they had specific heavy metal uptakes independent of the tailings treatment. Under the experimental conditions, Q. virginiana was highly tolerant to heavy metals. Variability in the root system development might explain why the Quercus spp. were more tolerant of tailings than were Salix spp. in this study. All species showed low BCF and TF values, except for the TF value for Zn and Cd of S. integra. Using short rotations, S. integra may be suitable for phytoextraction of tailings in temperate regions. Based on the present findings, Q. virginiana was the most metal-tolerant species at the seedling stage and is a potential candidate for phytostabilization of mine tailings in southern China.