1 Introduction

With the rapid progress of science and technology, lead–zinc mineral resources have been widely used in the electrical industry, machinery industry, military industry, biomedicine, and other fields, which are indispensable for the development of national economy (Li et al., 2014a; Mudd et al., 2017; Zhang et al., 2012). This has promoted the increase of mining intensity; as a result, numerous tailings were produced. It is worth to note that the soils of the tailings contain amounts of harmful heavy metals, such as copper (Cu), zinc (Zn), lead (Pb), chromium (Cr), and cadmium (Cd). These heavy metals might gradually migrate via food chains in the coming years, thus might cause varying degrees of hazards to the surrounding life forms (Gabarrón et al., 2018; Kan et al., 2021; Liu et al., 2017). More severely, some labile tailing ponds might fail; thus, heavy metals might flow into farmlands, rivers, or even villages, causing large-scale environmental pollutions (Agurto-Detzel et al., 2016; Jacobson & Faust, 2014). Therefore, it is crucial to remediate the tailing ponds and surrounding heavy metal-contaminated areas.

At present, remediation technologies for the heavy metal pollutions in soils mainly include physical remediation, chemical remediation, and bioremediation. The conventional physical and chemical remediation methods are expensive and again yield toxic by-products (Audu et al., 2020; Zhang et al., 2019). In contrast, the bioremediation strategy by utilizing microorganisms as the “removers” has advantages of being low-cost, simple-operational, and large-scale-applicable, thus has been conceived to be an effective alternative to remove the heavy metals from the contaminated soils (Mani and Kumar, 2014; Dixit et al., 2015; Akcil et al., 2015; Katiyar et al., 2020). Studies have shown that microorganisms usually carry out the repair process through 3 different pathways, namely, biosorption, bioaccumulation, and biotransformation. However, the biggest obstacle for the bioremediation technique is that this strategy needs to be performed based on specific environmental conditions. This is mainly due to that the diversity and relative abundance of tolerant microorganisms in different heavy metal-polluted areas are usually different (Oliveira and Pampulha, 2006; Li et al., 2018; Liang et al., 2018). Many previous studies have shown that the composition, nature, and concentration of heavy metals in the polluted areas would cause changes of microbial communities (Margesin et al., 2011; Zhang et al., 2016). For example, Lopez et al. (2017) found that the higher the chemically available Ni content in the soil, the more the relative abundance of Proteobacteria (especially Alphaproteobacteria) and Actinobacteria, and the less that of Chloroflexi. Further, the bacterial phyla such as Proteobacteria, Acidobacteria, and Bacteroidota usually are more abundant than others in the heavy metal-contaminated environments (An et al., 2018; Lin et al., 2019). Except for heavy metals, the soil physical–chemical properties, such as soil texture, water content, pH, and content of nutrients (e.g., organic carbon, total nitrogen, and total phosphorus), will also greatly affect the structure of the microbial communities (Cui et al., 2020; Lin et al., 2019). For example, Liu et al. (2014) reported that the soil pH value played a crucial role in determination of the microbial community structure, which was positively correlated with the abundance of the bacterial phylum Planctomycetes, the bacterial genera Blastomonas and Chloroflexus. Li et al. (2014b) also reported that the composition of the bacterial community in forest soil was significantly related to soil pH, TN, and TP. Together, it is conceivable that to obtain prominent bioremediation effect, the tolerant microorganisms in the contaminated environment should be selected first (Li et al., 2020), which means investigation of the microbial community structure is of great importance.

Traditional methods for determining the microbial community structures normally depend on laboratory culture or PCR-based techniques, which are inefficient and insensitive, and the results only reflect a limited portion of the entire microbial population (Chikere et al., 2019; Pan and Yu, 2011; Shi et al., 2002). In recent years, many research groups have demonstrated that high-throughput 16S and 18S rRNA amplicon sequencing is a more direct, accurate, and efficient method for analyzing the microbial community structure in contaminated samples (Abia et al., 2018; Kang et al., 2020; Xu et al., 2020). In this study, we chose two lead–zinc tailings (i.e., LJP and SP tailings) located in the Qinling Mountains, Ningqiang county of Shaanxi Province in northwestern China as the research objects. The LJP and SP tailings are from the same mine, i.e., Ningqing mine which has a mining history of 37 years. The compositions of the both tailing soils are mainly shales and phyllites, and the soils show characteristics of being loose, highly water-permeable, and poorly erosion-resistant. Further, vegetation coverages in the tailing ponds are extremely low (Zhou et al., 2020). Under these backgrounds, it seemed that the two tailings have been suffered from long-term and severe heavy metal pollutions. Specifically, in this work, 6 sites in the SP and LJP tailing ponds and one site in a surrounding farmland of the LJP tailing were selected for soil sample collections. The bacterial communities (including richness and diversity) in the soils were then assessed by 16S rRNA high-throughput sequencing method; and the heavy metal species and concentrations and the chemical properties such as pH, SOM, TP, TK, and TN of these soils were also analyzed by a series of methods. By comparison of the bacterial communities and the soil chemical properties of the 6 tailing soils and the control farmland soil, we aimed to reveal whether the bacterial communities had been changed by the heavy metal contaminations, and if so, how and to which extent they had been varied. There are few reports regarding the ecological systems of this area, and before this work, the microbial community structures of the SP and LJP lead–zinc tailings were completely unknown. Our results showed that due to various heavy metal pollutions, the bacterial community richness and diversity in the tailing soils have been greatly changed, and the relative abundances of certain bacterial species such as s__unclassified_g__Sulfurifustis, s__unclassified_f__Rhodanobacteraceae, s__unclassified_g__Conexibacter, s__unclassified_g__norank_f__norank_o__Gaiellales, and s__unclassified_g__Blastococcus have been increased, indicating that these bacteria had strong heavy metal-tolerant capacity. Together, we have initially investigated the heavy metal species/contents, the soil chemical properties, and their relationships with the indigenous microbial communities of these specific tailing soils, thus the findings of this work have provided a cornerstone for future bioremediation of these tailing ponds.

2 Materials and Methods

2.1 Soil Sample Collections

The soil samples were collected on December 21, 2019, and were from two lead–zinc tailing ponds in Liujiaping (LJP) and Shanping (SP) villages, which are both situated in the Qinling Mountains, Ningqiang county of Shaanxi Province in northwestern China (Fig. 1). The straight-line distance between the two sampling sites is 41 km. Specifically, for the LJP and SP tailing ponds, 3 sites located at different altitudes were chosen for sampling, respectively; in addition, soil from a farmland (NT) which is 101 m away from the LJP tailing pond was collected as well (Fig. 1), and it was used as a control for further analyses. The geographic characteristics of the 7 sampling sites are listed in Table 1. For each site, 3 points with a distance of approximately 5 m were randomly selected for sampling. The soil samples were collected and placed in sterilized containers. Each soil sample was divided into two portions, one portion was used for soil chemical properties and heavy metal contents analyses, and the other was used for bacterial community analysis.

Fig. 1
figure 1

Maps of the sampling sites. Upper-right panel, Shanping (SP) tailing; lower-right panel, Liujiaping (LJP) tailing and LJP farmland

Full size image
Table 1 Geographic information of the sampling sites
Full size table

2.2 Analyses of Chemical Properties and Heavy Metal Concentrations of the Soil Samples

Soil pH was measured using a pH meter (the mass to volume ratio of soil and water was 1:2.5) (Reijonen et al., 2016). Soil organic matter (SOM), total nitrogen (TN), total phosphorus (TP), and total potassium (TK) were respectively determined by potassium dichromate-sulfuric acid colorimetry, Kjeldahl method, alkali molybdenum anti-colorimetry, and alkali fusion-flame spectrophotometry (Song et al., 2019; Tsiknia et al., 2014). Ammonia nitrogen (NH4+_N) was measured by the phenol-hypochlorite colorimetric method (Li et al., 2014b); nitrate nitrogen (NO3−_N) was measured by ultraviolet spectrophotometry; and metallic elements (i.e., Cu, Zn, Cr, Pb and Cd) were measured by a flame atomic absorption spectrophotometer (TAS-990F, Beijing Purkinje General Instrument Co., Ltd., Beijing, China) (Zhang et al., 2016).

2.3 DNA Extraction and 16S rRNA Amplicon Sequencing

Genomic DNA of the soil samples were extracted by Power Soil DNA Isolation Kit (MoBio) following the manufacturer’s procedure. The extracted genomic DNA was used as a template for PCR amplification of the V4 region of the bacterial 16S rRNA gene. The primers used for PCR amplification were 515F and 806R (Cabanet al., 2018; Nottingham et al., 2018); the primer sequences were as follows: 515F: 5′GTGCCAGCMGCCGCGGTAA3′ and 806R: 5′GGACTACHVGGGTWTCTAAT3′. PCR amplification conditions were as follows: denaturation at 94 °C for 5 min, then 30 cycles of denaturation (94 °C/30 s), annealing (52 °C/30 s) and extension (72 °C/30 s), and final extension was performed at 72 °C for 10 min. PCR products (around 245–260 bp) were purified with EZNA Gel Extraction Kit (Omega, USA). Sequencing libraries were generated using NEBNext® Ultra™ DNA Library Prep Kit for Illumina® (New England Biolabs, USA) following the manufacturer’s protocols. The library quality was assessed on the Qubit@ 2.0 Fluorometer (Thermo Scientific) and Agilent Bioanalyzer 2100 system. The libraries were sequenced on an IlluminaHiseq2500 platform and 250-bp paired-end reads were generated (the sequencing was performed by Mega Biomedical Technology Co., Ltd., Shanghai, China).

2.4 Statistical Analysis

Initially, the raw sequencing reads were grouped into different operational classification units (OTUs), based on ≥ 97% sequence similarity by using USEARCH 7.0 (Kang et al., 2020). On the basis of the OTU clustering results, the following analyses were performed. The RDP classifier software (version 2.2) was used to categorize the above OTUs into their exact taxonomy, the threshold was set to 0.7, and the classification values below the threshold were defined as “unclassified” (Xu et al., 2020). The following bioinformatic analyses were performed based on the softwares presenting on Majorbio I-Sanger Cloud Platform (Majorbio Co., Ltd., Shanghai, China). In detail, the bacterial community richness (indexes: Sobs, Ace, and Chao1), diversity (Shannon and Simpson), and coverage rate indicating the percentages of the 16S rRNA sequences in the sample that have been tested were all analyzed by Mothur software (V 1.30.1) (Schloss et al., 2009). The R language tool was used to generate column charts which exhibited the bacterial phyla composition of each soil sample, and the significance of differences with respect to the relative abundance of the bacterial phyla in these soil samples was analyzed using Kruskal–Wallis rank sum test method. To cluster the soil samples, Principal coordinate analysis (PCoA) was performed based on the counts of OTUs on phylum level, and it was executed by the QIIME software (Jiang et al., 2020). Statistical PCoA difference analyses of the sampling sites were analyzed by Adonis analysis, and the Bray–Curtis distance-based method was used. Canonical correlation analysis (CCA) was used to assess the pluralistic relationships among the bacterial community diversities (the datasets used were the counts of OTUs on genus level), the soil chemical properties, and the soil heavy metal contents, which was accomplished by Canoco 5.0 software. Pearson correlation analysis was used to analyze the correlations between the relative abundance of 30 dominant bacterial phyla and the heavy metals, as well as the chemical properties of the soil samples, and the Origin software was used to draw the correlation heatmap. SPSS (version 24.0) was used to analyze the significant differences of the heavy metal contents, the chemical properties, the bacterial community richness and diversity (OTUs on phylum level), and the relative abundance of dominant bacterial phyla/species among the soil samples, in which statistically significant differences were considered at p < 0.05.

3 Results

3.1 Heavy Metal Concentrations of the LJP, SP Tailing Soils, and the Control Farmland Soil

On the whole, there were significant differences with respect to heavy metal concentrations of the soil samples collected from LJP, SP tailing ponds, and farmland (p < 0.01). The results are summarized in Table 2. It is obvious that except for Cu and Pb in SP_1 and Zn, Cr, and Pb in LJP_3, in other cases, the concentrations of the heavy metals in the tailing soils were significantly higher than those in the farmland soil (p < 0.01), especially that of Cd. The most severely contaminated site was SP_2, with the Cu and Zn concentrations top-ranked, and the Cr, Pb, and Cd concentrations secondly ranked. However, although the concentration of Cd in SP_2 was hundreds of times exceeding the standard level (i.e., 0.6 mg/kg), those of Cu, Zn, Cr, and Pb were still under standard thresholds, which are respectively 100, 300, 250, and 170 mg/kg according to the “Soil Environmental Quality and Agricultural Land Soil Pollution Risk Control Standard of China (GB15618-2018).” LJP_2 seemed to be the second polluted site, in which the concentrations of Cr and Pb were the highest among those of all the sites. The concentration of Pb in LJP_2 was 68 and 88 times higher than those of the farmland soil and the SP_1 soil, respectively. Compared with others, SP_3 and LJP_3 were both mildly contaminated. Overall, compared with the LJP tailing, the SP tailing had suffered more severely from Cd pollution. Further, although the Cd level in the farmland was the lowest among those of all the sites, it also exceeded the standard value, indicating that Cd had been brought into this farmland from the nearby tailings through farming or other unknown reasons. Together, the two tailing soils were strikingly polluted by Cd, while the other heavy metal pollutions were relatively moderate.

Table 2 Heavy metal concentrations of the soil samples collected from the LJP and SP tailings
Full size table

3.2 The Chemical Characteristics of the Soil Samples

Consistent with the results of the above heavy metal content changes, significant differences (p < 0.01) in the chemical properties of the soil samples were also detected (Table 3). Specifically, the SP_1 and SP_3 tailing soils were slightly alkaline, but the rest of the soil samples were all acidic, from which a lowest pH value of 3.97 ± 0.32 was recorded in the LJP_2 site. Except for pH, the concentrations of NH4+_N and NO3−_N of LJP_2 were also the highest among those from all the sites. The heavy metal levels of the SP_2 site were either the highest or the second-highest (Table 2), correspondingly, the SOM and TP contents in this site were both the highest, and TN, NH4+_N, and NO3−_N contents were second-placed among those of all the samples. Further, the average SOM, TN, TP, TK, NH4+_N, and NO3−_N values of the 3 SP sites were respectively 8.59, 0.55, 0.54 and 1.62 g/kg, and 7.77, 5.54 mg/kg, and those of the 3 LJP sites were respectively 3.71, 0.37, 0.18 and 1.41 g/kg, and 7.70, 4.56 mg/kg. Thus, the contents of SOM, TN, TP, NH4+_N, and NO3−_N were all ranked as SP > JLP > NT, and the content of TK was ranked as SP > NT > LJP.

Table 3 Chemical properties of the SP and LJP tailing soils
Full size table

3.3 Bacterial Community Richness and Diversity in the Soil Samples

A total of 5, 923 OTUs were obtained, following the high-throughput 16S rRNA sequencing and clustering analyses. To understand the bacterial community richness and diversity of the above soil samples, the OTUs were categorized into exact taxonomy followed by analysis with the Mothur software (see “Sect. 2”). The results are summarized in Table 4. In general, there were significant differences in both bacterial community richness and diversity among the 7 samples (p < 0.05). According to the values of the community richness indexes Sobs, Ace, and Chao1, it was inferable that the community richness of the 7 soil samples was trending as SP_3 > SP_1 > SP_2 > NT > LJP_1 > LJP_2 > LJP_3. The Shannon and Simpson indexes indicated that the variation trend of the community diversity was highly analogous to that of the community richness, projecting as SP_3 > SP_1 > SP_2 > LJP_3 > NT > LJP_1 > LJP_2. Therefore, compared with those of the farmland soil, the bacterial community richness and diversity in the tailing soils were either increased or decreased.

Table 4 Bacterial community richness and diversity in the soil samples
Full size table

3.4 Bacterial Community Compositions in the Soil Samples

A total of 703,464 valid sequences were detected from the 21 soil samples, including 5923 OTUs (similarity ≥ 97%), 40 phyla, 119 classes, 283 orders, 459 families, 902 genera, and 1914 species. The bacterial community compositions and variations in the soil samples were then analyzed on the phylum level. The average relative abundances of the dominant bacterial phyla which mainly included Actinobacteriota, Acidobacteriota, Chloroflexi, Cyanobacteria, Gemmatimonadota, WPS-2, Bacteroidota, and Firmicutes were significantly different (p < 0.05) among the 7 sampling sites (Fig. 2A). Specifically, compared with those of the control farmland soil, the average relative abundances of Actinobacteriota, Acidobacteriota, Cyanobacteria, Gemmatimonadota, WPS-2, Bacteroidota, and Firmicutes in certain contaminated LJP and SP soils (e.g., SP_1, SP_2, SP_3, and LJP_2) were significantly increased (Fig. 2B; Table S1). For example, compared with that of 19.75% ± 2.63% in the control soil, the average relative abundance of Proteobacteria in the LJP_1 site has been increased to 31.52% ± 14.19%. On average, in LJP_2, Acidobacteriota accounted for 25.42% ± 13.98% of the total bacterial community, in contrast, only 3.62% ± 2.80% of Acidobacteriota were detected in the control soil. Further, the average relative abundance of Actinobacteriota of the SP_1 tailing soil reached 30.72% ± 3.58%, which was in significant contrast with that of 15.93% ± 2.82% in the control soil (Table S1). By comparison with the above bacterial phyla, Chloroflexi was on a downswing trend in the contaminated soils, especially in the SP soils, in which a lowest value of 8.12% ± 3.20% was detected in SP_2 (43.13% ± 5.43% in the control soil) (Table S1). Further, on the species level, the average relative abundance of certain species especially those belonging to the Proteobacteria, Actinobacteriota, and Acidobacteriota phyla in the contaminated LJP and/or SP soils was found to be significantly increased (Fig. S1; Table S2). These species mainly included s__unclassified_g__Sulfurifustis and s__unclassified_g__Conexibacter (respectively 5.87% ± 2.89% and 5.44% ± 2.37% found in LJP_1 compared with respectively 0.00% ± 0.00% and 0.17% ± 0.03% found in NT); s__unclassified_f__Comamonadaceae and s__unclassified_g__Blastococcus (respectively 3.14% ± 2.17% and 3.63% ± 2.88% found in LJP_2 compared with respectively 1.59% ± 0.66% and 1.14% ± 0.23% found in NT); s__unclassified_f__Rhodanobacteraceae, s__uncultured_bacterium_g__Metallibacterium, s__uncultured_bacterium_g__Acidiphilium, and s__unclassified_g__norank_f__norank_o__Gaiellales (respectively 6.69% ± 1.01%, 4.17% ± 2.62%, 4.38% ± 1.77%, and 4.84% ± 4.07% found in SP_1 compared with respectively 0.00% ± 0.00% and 0.04% ± 0.00% found in NT); and s__uncultured_forest_soil_bacterium_g__norank and s__uncultured_forest_soil_bacterium_g__norank_f__norank_o__Subgroup_13 (respectively 6.09% ± 1.13% and 17.79% ± 13.50% found in SP_2 compared with respectively 0.00% ± 0.00% found in NT) (Table S2).

Fig. 2
figure 2

Average relative abundance and predominant bacterial phyla in the 7 soil samples. A Kruskal–Wallis H test bar plot showing the significance of differences with respect to the relative abundance of the 10 dominant bacterial phyla among the 7 soil samples. In the vertical axis, dominant bacterial phyla are shown; the column length represents the average relative abundance of the corresponding phylum; p values are shown on the right side, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. B Bar plot visually showing the average relative abundances of the predominant bacterial phyla in the 7 soil samples

Full size image

3.5 Bacterial Community Variations Among the Soil Samples

To better visually assess the bacterial community variations among the tailing soils and the control farmland soil, Principal coordinate analysis (PCoA) on the phylum level was performed (see “Sect. 2”). Together, the two main selected coordinate axes, i.e., PC1 and PC2, explained 54.76% of the total variation of the bacterial community structure (Fig. 3). Compared with that of the control farmland soil, the eclipse matching areas of the LJP_1, LJP_2, LJP_3, and SP_1 sites were apparently larger, indicating that the bacterial communities in these sites were varied to a higher degree than that of the farmland soils (Fig. 3). Consistently, the Adonis analysis results showed that all the bacterial communities of the 6 tailing soils were significantly different from that of the control farmland soil, in which a R2 value of 0.7135 and a p value of 0.001 were detected (Fig. 3).

Fig. 3
figure 3

Principal coordinate analysis (PCoA) at the phylum level. The Y-axis and X-axis represent the two main coordinate axes, and the percentage values of the coordinate axes represent interpretations of the differences in sample composition. The closer the two sample points are, the more similar their bacterial phylum composition

Full size image

3.6 Correlations Between the Bacterial Community Diversity and the Environmental Parameters

To understand the impact of the individual environmental factor on the relative abundance of the dominant bacterial phyla, a Pearson heatmap correlation analysis was performed. The results demonstrated that the relative abundance of most of the dominant bacterial phyla showed different degrees of significant correlations with the heavy metals and/or the soil chemical properties (Fig. 4). Specifically, Acidobacteriota were significantly positively correlated with NO3−_N, Cr, and Pb. Cyanobacteria were significantly positively correlated with TP, Cu, and Zn. Actinobacteriota were significantly negatively correlated with Pb and significantly positively correlated with pH, TP, SOM, TN, and Cd. Similar to Actinobacteriota, Bacteroidota were also significantly positively correlated with TP, SOM, TN, and Cd and additionally with Zn. In contrast to that of Actinobacteriota and Bacteroidota, the relative abundance of Chloroflexi was negatively correlated with most of the environmental parameters and was significantly downregulated by TP, SOM, TN, Cd, and NO3−_N. Further, Gemmatimonadota and Myxococcota were both significantly negatively correlated with Cr, besides, Myxococcota were also negatively correlated with NH4+_N and Pb (Fig. 4).

Fig. 4
figure 4

Pearson heatmap correlation analysis of the relative abundances of 30 bacterial phyla and their correlation significances with the environmental factors. Hierarchical clustering on the side: the correlations of the bacterial taxa, hierarchical clustering on the top: classified relationships of the environmental parameters. Legend indicator: the correlation coefficients between the relative abundances of the bacterial phyla and the environmental parameters; warm tone represents positive correlation; cold tone represents negative correlation; *0.01 < p ≤ 0.05, **0.001 < p ≤ 0.01, ***p ≤ 0.001

Full size image

Further, to uncover to which extent that each specific environmental factor had changed the bacterial community diversity, canonical correlation analysis (CCA) on the genus level was performed. In Fig. 5, the first main selected axis (CCA1) explained 22.71% of the bacterial community diversity, and the second main selected axis (CCA2) explained 15.73% of this value. Specifically, the bacterial community diversity of SP_1, SP_2, and SP_3 were all positively correlated with TK, pH, TP, TN, SOM, NO3−_N, Cd, and Zn and were all negatively correlated with Cu, NH4+_N, Pb, and Cr. The relationships between the bacterial community diversity and the environmental factors in the LJP_1 and LJP_2 soils were nearly contrary to that of the SP samples, i.e., they were positively correlated with NH4+_N, Cu, Pb, and Cr, but negatively correlated with TK, pH, TP, TN, Cd, SOM, Zn, and NO3−_N. Further, LJP_3 seemed to be the most similar site to the control site, as in both situations, the bacterial community diversities were only mildly positively correlated with TK and negatively correlated with all the other environmental factors (Fig. 5). Moreover, TP and Cd were the dominant positive factors affecting the bacterial community diversities of SP_1 and SP_3; TP and pH were the main positive factors affecting the bacterial community of SP_2; and Cr was the core negative factor for the bacterial community diversities in all the 3 SP sites. In a striking contrast to those of the SP sites, the main positive factors of the bacterial community diversities of the LJP_1 and LJP_2 sites were Cr and Pb, while the main negative factors of which were TP, Cd, and pH (Fig. 5).

Fig. 5
figure 5

Canonical correlation analysis (CCA) demonstrating the correlations between the bacterial community diversities and the environmental parameters. Red-colored arrows: the environmental factors; arrow length represents the extent of influence; the angle of deviation between the sample site and the environmental parameters indicates positive or negative correlation, i.e., acute angle indicates positive correlation; right angle indicates no correlation; and obtuse angle indicates negative correlation

Full size image

4 Discussion

In recent years, Ningqing County, Shaanxi Province has produced a large number of tailing ponds due to frequent mining activities. However, there are few reports regarding the ecological system of the tailings in this area, which hinders the process of bioremediation. In this study, we have conducted an initial analysis of the bacterial communities of the two main lead–zinc tailings (i.e., LJP and SP tailings) located in this area. Our results showed that due to the heavy metal pollutions, the chemical properties and the bacterial community structures of the tailing soils have been changed to various degrees.

4.1 Mutual Influence Relationships Among the Soil Heavy Metal Contents, Chemical Properties, and Indigenous Ecosystems

The LJP and SP tailing ponds were geographically closed; however, they contained significant different heavy metal contents (Table 2). This is mainly due to that the heavy metal accumulation years of the two tailings were different; further, long-term weathering, rainwater leaching, and soil adsorption might also have changed the heavy metal contents. Correspondingly, the chemical properties of the two tailing soils were also significantly different (Table 3). The average SOM, TN, TP, and TK contents were much higher in the SP tailing than those in the LJP tailing, and this variation trend was resembled to that of the Cd content. Thus, it is presumable that the Cd contamination had a crucial role in determining these soil chemical properties. Further, except for TK, the values of all the chemical parameters of the control farmland soil were placed between those of the SP and LJP tailing soils (Table 3); together with the results of the heavy metal analyses (Table 2), it seemed that different levels of heavy metal contaminations might have changed the soil chemical properties to different extents. One possible approach is that the heavy metals might have changed the indigenous soil ecosystems, which would indirectly cause changes of the production and consumption of the soil nutrients. For instance, Liu et al. (2014) showed that with the stress of heavy metals, tolerant microbes such as Acidobacteria which were capable of producing acid were enriched, and thus, the soil pH was lowered. Our results were in high agreement with this assumption, as the LJP_2 site had the highest relative abundance of Acidobacteria (Table S1), and the lowest pH level among those of all the samples (Table 3). However, on the other hand, the soil chemical properties might in turn affect the heavy metal contents. Several studies have shown that the concentrations of the heavy metals would be decreased along with decreasing of the soil pH, and this phenomenon was due to that abundant organic acids were existing in these acidic soils which could promote dissolution or reduction of the heavy metals (Ash et al., 2016; Onireti et al., 2017). It has also been repeatedly demonstrated that the heavy metal contents were strongly correlated with the soil SOM and TOC contents (Hooda & Alloway, 1998; Strawn and Sparks, 2000; Šípková et al., 2013), since these soil organic carbon contents tended to affect the amount of soil agglomerates which could directly absorb heavy metals (Choudhury et al., 2014). Together, it seemed that as the situations found in the previous studies (Ash et al., 2016; Gao et al., 2021; Onireti et al., 2017; Šípková et al., 2013), in our case, mutual influence relationships were also found to be existing among the heavy metal contents, the soil chemical properties, and the indigenous soil ecosystems in these tailing soils.

4.2 Heavy Metals in the Tailing Soils Have Changed the Bacterial Community Structures

Environmental factors such as pH, SOM, TN, TP, and TK are considered to be the main driving force of the microbial community changes (Tilman and Kareiva, 2018; Lin et al., 2019; Gao et al., 2021). Normally if these factors were constant, the microbial community structures would be kept steady. As above analyzed, the heavy metal contamination may change the soil chemical properties; it is conceivable that this kind of alteration will be concomitant with alterations of microbial community structures. In the present work, the heavy metal concentrations, soil chemical properties, and bacterial community structures of the LJP and SP tailing soils were respectively analyzed, as expected, it was revealed that compared with that of the control farmland soil, the bacterial community structures in these heavy metal-contaminated soils were significantly changed (Table 4, S1; Fig. 2, 3, 4, 5). Specifically, the bacterial community richness and diversity in the contaminated soils were either increased or decreased, and compared with those of the control soil, the average relative abundance of certain bacterial genera especially those belonging to the Actinobacteria and Proteobacteria phyla in the contaminated LJP and SP soils was notably increased (Fig. S2). These bacterial genera mainly included Gaiella (Actinobacteria), Arthrobacter (Actinobacteria), Pseudomonas (Actinobacteria), Ramlibacter (Proteobacteria), Bradyrhizobium (Proteobacteria), Metallibacterium (Proteobacteria), Massilia (Proteobacteria), unclassified_f__Acetobacteraceae (Proteobacteria), and norank_f__Xanthobacteraceae (Proteobacteria) (Fig. S2). For example, compared with those of 0.16% and 0.12% in the control soil, the average relative abundances of Gaiella and norank_f__Xanthobacteraceae in the SP_1 site have been respectively increased to 4.49% and 1.97%. On average, in SP_2, Arthrobacter accounted for 1.44% of the total bacterial community, in contrast, only 0.05% of which was detected in the control soil. The average relative abundances of Bradyrhizobium and unclassified_f__Acetobacteraceae have respectively reached 1.79% and 2.16% in the LJP_1 soil, in contrast, only 0.47% and 0.06% of which were respectively recorded in the control soil (Fig. S2). This result is highly consistent with the situations found in some other polluted areas, i.e., several previous studies have pointed out that the bacterial genera such as Gaiella, Arthrobacter, Bradyrhizobium, unclassified_f__Acetobacteraceae, and norank_f__Xanthobacteraceae were dominant genera in the heavy metal-polluted soils (Bier et al., 2015; Gao et al., 2021; Guo et al., 2019; Osman et al., 2021). Thus, it was speculated that these bacteria genera were heavy metal-resistant, which means under the stimulation of heavy metals, they may alleviate the toxic effects rendered by the heavy metals, probably through their own structural characteristics and/or secondary metabolic activities, thereby adapting to the contaminated environment and becoming the dominant bacteria (Vig et al., 2003; Liu et al., 2014; Lopez et al., 2017; Becerra-Castro et al., 2015; Wang et al., 2010). All the species of Proteobacteria are gram-negative bacteria which contain a layer of lipopolysaccharide surrounding the cell wall. Studies have shown that certain Proteobacteria such as Bradyrhizobium and unclassified_f__Acetobacteraceae were capable of preventing heavy metals from entering into the cell cytoplasm through this lipopolysaccharide layer (Pereira et al., 2006). Other studies have also pointed out that the members (e.g., Gaiella and Arthrobacter) in the Actinobacteriota phyla usually possess a variety of physiological and metabolic properties, such as the production of extracellular enzymes and the formation of multiple secondary metabolites. These secondary metabolites and enzymes could interact with each other thereby removing pesticides, heavy metals, and other foreign biological compounds (Alvarez et al., 2017; Ashok and Geetha, 2018). Bacteroidota are capable of synthesizing all necessary organic compounds from carbon dioxide, i.e., they are chemoautotrophic bacteria, thus can easily survive under heavy metal-polluted environments (Li et al., 2016; Mehrani et al., 2020). The species in Acidobacteriota can convert complex organic carbon into short-chain fatty acids, which may promote dissolution or absorption of heavy metals by hyperaccumulator (Eichorst et al., 2007; Fierer et al., 2007); thus, they are normally one of the most heavy metal-tolerant bacteria. Certain previous studies have also found that Chloroflexi can well survive in heavy metal-contaminated soils, as that this kind of bacteria is capable of growing through various nutritional pathways such as photosynthetic, heterotrophic, photoautotrophic, and chemoautotrophic (Fernández et al., 2008). However, in this study, the relative abundance of Chloroflexi in the two tailing soils was not significantly increased. One possible explanation for this phenomenon is that the concentration of the heavy metals in these locations has far exceeded the tolerance of Chloroflexi, leading to extinction of species in this phylum. Nevertheless, our results showed that heavy metals can cause changes of the bacterial community structures, which was consistent with the results from most of the other studies (Gao et al., 2021).

4.3 The Relationships Between the Environmental Parameters and the Bacterial Community Structures

Among the contaminated soils, the community-structure similarities reflecting the environmental changes were most significant between LJP_1 and LJP_3, and SP_1 and SP_3, as intersecting matching areas were respectively detected (Fig. 3). In general, the bacterial community structures were more widely varied in the LJP tailing soil than those in the SP tailing soil (Fig. 3). This result echoed the results with respect to soil heavy metal contents and chemical properties (Table 2, 3), which also displayed larger changing scopes of the LJP_2 and LJP_3 sites. Further, Pearson heatmap correlation and Canonical correlation (CCA) analysis results showed that the heavy metals such as Cd, Pb, and Cr and the soil chemical indexes such as pH, SOM, TN, TP, and TK were strongly correlated with the relative abundance and diversities of the bacterial communities (Fig. 4, 5). These results indicated that the heavy metals and soil chemical characteristics could both alter the bacteria communities in the tailing soils, which were in high agreement with the results from previous studies. For instance, Lin et al. (2019) have studied the changes of microbial community structures under different heavy metal contamination levels in paddy soils and showed that most dominant bacteria such as Armatimonadetes, Chloroflexi, Verrucomicrobia, and Planctomycetes were in significant negative correlation with soil Cd content, especially in the moderate contamination region. An et al. (2018) have investigated the microbial community structures in Pb-contaminated Lou soils and found that the relative abundances of Nitrospirae, Gemmatimonadetes, and Planctomycetes were negatively correlated with the Pb concentration. pH has been repeatedly proved to be an important determinant for the bacterial communities in the heavy metal-contaminated soils (Jiang et al., 2021; Kuang et al., 2013; Li et al., 2014b), and positive correlations between pH and the abundance of certain bacterial genera such as Fusarium, Blastomonas, and Chloroflexus have been recorded (Liu et al., 2014). In addition to pH, other soil chemical factors, e.g., SOM, TN, TP, and TK, were also frequently shown to be significantly related to the bacterial community structures. For example, SOM seemed to be an particularly essential environmental element that could influence the relative abundance of a variety of microbes, such as the bacterial phyla Latescibacteria, Ignavibacteriae, Chlorobi, and Nitrospirae (Lin et al., 2019). Further, Zhao et al. (2019) have studied the microbial community of long-term heavy metal-polluted lands and found that the relative abundances of Proteobacteria and Firmicutes were negatively associated with the soil available potassium (AK). Our results also showed that the influence of heavy metals and soil chemical properties on the bacterial community structure were crucial (Fig. 4, 5). Generally, the positive correlation traits of these factors were more significant than the negative ones, especially when reflected the relative abundances of the dominant bacterial phyla such as Actinobacteriota, Bacteroidota, Chloroflexi, and Cyanobacteria. This result was highly resembled to those found in many previous studies, indicating that these bacteria are indeed strongly tolerant to heavy metal contaminations. Thus, they have outgrown in the whole bacterial population in the tailing soils.

In conclusion, the present work in this manuscript showed that: (1) the LJP and SP tailing ponds were both polluted by heavy metals, especially by Cd; however, the heavy metal contents and chemical properties in these two tailing soils were significantly different. (2) Correspondingly, significant differences in the bacterial community richness and diversity were recorded in the two tailing soils, which were demonstrated as SP > LJP; the result was consistent with the facts that the SP tailing soil contained higher level of Cd, and that LJP possessed more diverse soil chemical properties. (3) Compared with that in the control farmland soil, the relative abundance of certain bacterial phyla in the contaminated LJP and SP soils have been increased or decreased significantly. Specifically, the relative abundances of Actinobacteriota and Bacteroidota were positively correlated with Cd, TP, TK, SOM, TN, and pH; Cyanobacteria were significantly upregulated by Cu, Zn, and TP; Acidobacteriota were notably positively correlated with Cr, Pb, and NO3_N; on the contrary, the relative abundance of Chloroflexi was negatively correlated with all the environmental parameters, especially with Cd, NO3−_N, TP, SOM, and TN. Together, the above results indicated that the heavy metals and the chemical properties of the tailing soils have jointly affected the bacterial community structures. Overall, in this manuscript, we have initially investigated the ecological systems of the SP and LJP tailing soils, and from which we successfully identified certain heavy metal-tolerant bacterial species respectively belonging to the Proteobacteria and Actinobacteriota phyla, e.g., s__unclassified_g__Sulfurifustis, s__unclassified_f__Rhodanobacteraceae, s__unclassified_g__Conexibacter, s__unclassified_g__norank_f__norank_o__Gaiellales, and s__unclassified_g__Blastococcus.