Effects of Pb Smelting on the Soil Bacterial Community near a Secondary Lead Plant

Secondary lead smelting is a widespread industrial activity which has exacerbated Pb or Cd contamination of soil and water across the world. Soil physicochemical properties, soil enzyme activities, heavy metal concentrations, and bacterial diversity near a secondary lead plant in Xuzhou, China were examined in this study. The results showed that secondary lead smelting activities influenced nearby soils. Soil acidification decreased one order of magnitude, with a mean value of 7.3. Soil organic matter also showed a downward trend, while potassium and nitrogen appeared to accumulate. Soil urease and protease activity increased in samples with greater heavy metal pollution, but overall the soil microbial biodiversity decreased. Soil heavy metal concentration—especially Pb and Cd—greatly exceeded the concentrations of Chinese Environmental Quality Standard for Soils (GB 15618-1995). Some environmental factors—such as pH, organic matter, enzyme activity, and the concentration of heavy metals—significantly affected bacterial diversity: compared with the control site, the Chao1 estimator decreased about 50%, while the Shannon diversity index dropped approximately 20%. Moreover, some genera have significant relationships with heavy metal concentration—such as Ramlibacter with Zn and Steroidobacter with Cd—which might act as bio-indicators for soil remediation. These results will provide a new insight in the future for reclaiming soil contaminants caused by secondary lead smelting.


Introduction
Lead is an indispensable component in anti-corrosives, anti-radiation materials, and lead-acid batteries, which has played an essential role in industrial circles [1,2]. In the past decade, global lead consumption has continued to increase [3]. Due to non-renewability [4], secondary lead has been an important source [5,6]. The worldwide production of refined lead was approximately 108.9 million tonnes in 2015; more than half, about 61.7 million tonnes, came from secondary lead; and 70% of secondary lead was disassembled from waste lead-acid batteries [3,7]. Unfortunately, the disassembling processes of waste lead-acid batteries-which contains crushing, fusion, reduction, and refining-causes large quantities of smoke dust and wastewater into the surrounding soil and water bodies [8][9][10][11]. These wastes can destroy ecosystems and threaten public health [12,13].
Due to the effect of Pb on the brain and cognition of children with resulting behavioral changes, lead pollution has widely been a concern [14][15][16][17]. Lead is a toxic heavy metal that has adverse effects on human lung, kidney, and reproductive and cardiovascular systems [13]. People are

Soil Processing and Pollution Assessment
Soil pH values were measured using a glass electrode (PHC-3C, Shanghai Leici, Shanghai, China) in a 1:5 suspension of ultrapure water [29]. Soil organic matter content was measured with the potassium chromate oxidation-colorimetric method [34]. The available phosphorus contents were determined by the hydrochloric acid ammonium chloride method [35]. The ammonium acetate extraction-colorimetric method was used to measure available potassium [33]. Soil nitrate nitrogen contents were determined by the calcium chloride-colorimetry method [36].
Soil urease activities were determined by the phenol colorimetry of sodium hypochlorite [37]. The soil triphenyl tetrazolium chloride method was used to measure soil dehydrogenase activity [37]. Protease activities were measured by the ninhydrin colorimetric method [37]. The phenol colorimetric method was used to measure soil polyphenol oxidase [37]. The fluorescein diacetate hydrolysis activities were determined by the fluorescein colorimetric method [37].
We calculated the pollution load index (PLI) to evaluate heavy metal contamination at each site, using Equations (1) and (2): In Equation (1), CF refers to the contamination factor; m contains Cu, Cd, Pb, Zn, and Cr, respectively, while i represents each study site. In Equation (2), when PLI is <1, there is no pollution; when PLI is ≥1, but <2, there is low pollution; when PLI is ≥2, but <3, there is medium pollution; when PLI is ≥3, there is high pollution [38]. According to the Chinese Environmental Quality Standard for Soils (GB 15618-1995), the acceptable contamination levels (mg·kg −1 ) are: Cu 35, Cd 0.2, Pb 35, Zn 100, Cr 90 [39].

Soil Processing and Pollution Assessment
Soil pH values were measured using a glass electrode (PHC-3C, Shanghai Leici, Shanghai, China) in a 1:5 suspension of ultrapure water [29]. Soil organic matter content was measured with the potassium chromate oxidation-colorimetric method [34]. The available phosphorus contents were determined by the hydrochloric acid ammonium chloride method [35]. The ammonium acetate extraction-colorimetric method was used to measure available potassium [33]. Soil nitrate nitrogen contents were determined by the calcium chloride-colorimetry method [36].
Soil urease activities were determined by the phenol colorimetry of sodium hypochlorite [37]. The soil triphenyl tetrazolium chloride method was used to measure soil dehydrogenase activity [37]. Protease activities were measured by the ninhydrin colorimetric method [37]. The phenol colorimetric method was used to measure soil polyphenol oxidase [37]. The fluorescein diacetate hydrolysis activities were determined by the fluorescein colorimetric method [37].
We calculated the pollution load index (PLI) to evaluate heavy metal contamination at each site, using Equations (1) and (2): Concentration o f each heavy metal Natural background value (1) In Equation (1), CF refers to the contamination factor; m contains Cu, Cd, Pb, Zn, and Cr, respectively, while i represents each study site. In Equation (2), when PLI is <1, there is no pollution; when PLI is ≥1, but <2, there is low pollution; when PLI is ≥2, but <3, there is medium pollution; when PLI is ≥3, there is high pollution [38]. According to the Chinese Environmental Quality Standard for Soils (GB 15618-1995), the acceptable contamination levels (mg·kg −1 ) are: Cu 35, Cd 0.2, Pb 35, Zn 100, Cr 90 [39].

DNA Extraction, PCR Amplification, and Illumina MiSeq Sequencing
According to the manufacturer's instructions, we used the FastDNA™ SPIN Kit for Soil (MP 112 Biomedicals, Solon, OH, USA) to extract DNA from 0.5 g fresh soil samples. The quality of the DNA extracted was examined by 1% (w/v) agarose gel electrophoresis. Additionally, the concentrations were measured by using a UV-vis spectrophotometer (NanoDrop 2000, Thermoscientific, Waltham, MA, USA). The V3-V4 regions of 16S rRNA genes of the bacteria were amplified using the primer set 338F (5 -ACTCCTACGGGAGGCAGCAG-3 ) and 806R (5 -GGACTACHVGGGTWTCTAAT-3 ). Amplification conditions are as follows: 2 min at 95 • C, 30 cycles of 95 • C for 30 s, 55 • C for 30 s, 72 • C for 30 s, and a final elongation for 5 min in a GeneAmp ® 9700 Thermo Cycler (ABI, Foster City, CA, USA). The triplicate amplicons were pooled together, electrophoresed on a 2% (w/v) agarose gel, and recovered through an AxyPrep DNA Gel Extraction Kit (AXYGEN, Hangzhou, China). This purified amplicon was quantified using a QuantiFluor TM -ST Fluorometer (Promega, Madison, WI, USA). Then a composite sequencing library was constructed by combining equimolar ratios of amplicons from all the 17 samples. Finally, the sample libraries were analyzed by using the Illumina Miseq platform (Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China) [35].

Bioinformatics and Statistical Analysis
After Illumina MiSeq sequencing, the original double-ended sequencing data was obtained, which needed to be controlled and filtered by Trimmomatic [40]. After extracting the non-repetitive sequences and removing the single unduplicated sequences, operational taxonomic units (OTU) were clustered at 97% similarity using USEARCH [41]. RDP classifier [42] was used to classify the species. These species' abundance bar plots were drawn according to the taxonomic results. Mothur software (University of Michigan, Ann Arbor, MI, USA) [43] was used to calculate the Chao1 estimator [44] and the Shannon diversity index [45]. Chao1 (the Chao1 estimator) and Shannon (the Shannon diversity index) reflect the richness and uniformity of a single sample (a specific region or ecosystem) community. The beta diversity distance matrix was calculated using Qiime based on the Bray-Curtis algorithm [41], which is the ratio between regional and local species diversity [46]. The non-metric multidimensional scaling (NMDS) index was analyzed by using the R vegan package, whose aim is to collapse information from multiple dimensions into just a few for better visualized and interpreted [47]. The redundancy analysis (RDA) was analyzed with the R vegan package to detect the relationship between environmental factors, samples, and bacterial flora [35]. The heatmap was used to explore the Pearson relationship between the bacterial community and heavy metals with the R pheatmap package [48]. Duncan's test was performed by SPSS 22 software (IBM, Shanghai, China), which were applied in the case of multiple comparisons [35]. Charts and graphs were performed using Origin 9.1 (OriginLab, Northampton, MA, USA) or the R project (R Development Core Team, Vienna, Austria) [49].

Effect of Secondary Lead Smelting on Soil Physicochemical Properties and Enzyme Activities
Soil physicochemical properties are fundamental indicators of soil quality (Table 1). Soil pH values ranged from 6.89 to 7.72, with a mean value of 7.30. Soil pH values near the plant were lower than the control group, indicating that the soil has a certain degree of acidification. Soil organic matter was between 1.06% and 4.11% with a mean value of 2.37%. All 16 samples were lower than the control group. The contents of available potassium ranged from 141.77 to 249.44 mg·kg −1 with the mean value of 202.26 mg·kg −1 , which was higher than the control group. Furthermore, the average concentration of available potassium was high enough to support seasonal crops. The contents of available phosphorus ranged from 4.50 to 58.37 mg·kg −1 with an average of 14.58 mg·kg −1 . While the distribution of available phosphorus had a wide span, it did not show regularity. The mean value of soil nitrate nitrogen was 4.94 mg·kg −1 with a range from 1.04 to 13.58 mg·kg −1 , which was higher than the control. However, although the nitrogen content differed from site to site, each site was enriched in nitrogen such that it provided sufficient concentration for plant growth (Table 1).
Five enzyme activities surrounding the secondary lead plant were determined (Table 1). Soil urease is the critical enzyme in the conversion of nitrogen in the soil and can be measured as a proxy for the status of soil nitrogen. It ranged from 0.3 to 0.61 mg·g −1 ·d −1 , which presented a higher activity than the control group, 0.27 mg·g −1 ·d −1 . Soil protease activity ranged from 0.64 and 0.91 mg·g −1 ·d −1 , which was much higher than the control, 0.18 mg·g −1 ·d −1 . Soil dehydrogenase, which exercises tight control on respiratory metabolism, ranged from 0.48 to 1.05 g·g −1 ·h −1 . Soil FDA hydrolytic enzyme, which is widely used in soil quality assessments, ranged from 0.12 to 0.80 mg·kg −1 ·h −1 . Finally, soil polyphenol oxidase enzyme, which could promote oxidation of phenols, ranged from 0.01 to 0.29 mg·g −1 ·h −1 . None of the 16 sampling sites around the plant showed any difference from the control soil for these three enzymes (Table 1). Note: a SOM: soil organic matter (%); b AK: available potassium (mg·kg −1 ); c AP: available phosphorus (mg·kg −1 ); d NO 3 − -N: soil nitrate nitrogen (mg·kg −1 ); e Urease: soil urease activities (mg·g −1 ·d −1 ); f Protease: soil protease activities (mg·g −1 ·d −1 ); g Dehydrolase: soil dehydrolase activities (g·g −1 ·h −1 ); h FDA: soil FDA dehydrogenase activities (g·kg −1 ·h −1 ); i Polyphenol oxidase: soil polyphenol oxidase activities (mg·g −1 ·h −1 ); j C: control site; Values in the same column followed by different small alphabetical letters are significantly different at p < 0.05 level based on Duncan analysis (n = 3).

Soil Heavy Metal Concentrations and Containment Assessment near the Secondary Lead Plant
Soil heavy metal contents were shown in Table 2. The concentration of all five heavy metals exceeded the acceptable contamination levels [39]. The heavy metal that occupied the first place was Cd with a peak concentration of 6.97 mg·kg −1 at S2, which was 24 times than its background limit. Pb was in the second place with a peak concentration of 223.76 mg·kg −1 at S15, surpassing its background value 13 times. Even at their lowest concentrations, Cd and Pb were nearly five times more abundant than their corresponding background values. Cu was the third most abundant heavy metal with peak concentrations of 127.03 mg·kg −1 and 123.13 mg·kg −1 at S13 and S14, respectively; both of which exceeded its background value 10 times. The concentrations of Zn and Cr did not exceed their background values. The highest concentration of Zn was 203.66 mg·kg −1 in S16, while Cr was 153.93 mg·kg −1 in S11.
Furthermore, the pollution load index (PLI), used for the detailed evaluation of the degree of soil heavy metal pollution surrounding the secondary lead plant, changed from 1.87 (S8) to 3.69 (S15), with a mean value of 2.69. According to the PLI, 2 sites (S6, S8) showed low heavy metal pollution, 12 sites (S1-S5, S7, S10-S13, S16) showed medium heavy metal pollution, while 4 sites (S9, S14, S15) showed high heavy metal pollution. Note: a PLI: pollution load index; Values in the same column followed by different small alphabetical letters are significantly different at the p < 0.05 level based on Duncan analysis (n = 3).

Effect of Secondary Lead Smelting on Soil Bacterial Community Diversity
Soil bacterial community diversity was assessed using high-throughput DNA sequencing technology. A total of 591,763 sequences were obtained from 16 soil samples. The length distributions of trimmed sequences ranged from 401 to 600 bp. Under the OTU (operational taxonomic units) at 97% similarity, all rarefaction curves were apt to approximate the saturation platform, which indicated the sequence reads were reasonable [50].
The Chao1 estimator and the Shannon diversity index are shown in Table 3. The Chao1 estimator was at the peak with a maximum index of 2484.27 in S15, and at a minimum of 1763.92 in S1. The Shannon diversity index varied from 6.02 to 6.68. However, the two indices were both lower than the control group. Compared with the control site, the Chao1 estimator decreased about 50%, while the Shannon diversity index dropped approximately 20%. 1984.25bc 6.18abc S10 2118.81bcd 6.23bcd S11 2194.34cde 6.29bcde S12 2305.61def 6.46efgh S13 1909.72ab 6.14ab S14 2043.28bc 6.53fghi S15 2484.26g 6.68i S16 2314.3def 6.38cdef Note: Values in the same column followed by different small alphabetical letters are significantly different at p < 0.05 level based on Duncan analysis (n = 3).
The non-metric multidimensional scaling (NMDS) index is shown in Figure 2. According to the species information contained in the sample, it is reflected in the multi-dimensional space in the form of points, and the degree of difference between different samples is reflected by the distance between the points and the point, and finally obtains the spatial location map of the sample [51]. The control group and the surrounding sampling sites of the secondary lead plant had a far distance, which announced that the secondary lead smelting activity had severe effects that reduced the soil bacterial diversity in the surrounding soil.
To further investigate the composition of the bacterial community's structure, a bar plot of the taxonomic distribution of phyla in soil samples was generated (Figure 3). In the surrounding area of the secondary lead plant, Proteobacteria, Acidobacteria, Chloroflexi, Actinobacteria, Gemmatimonadetes, Nitrospirae, Bacteroidetes, Latrscibacteria, Planctomycetes, Verrucomicrobia, Saccharibacteria, Firmicutes, GAL15, and WS6 occupied more than 90% of the total sequences of each soil sample. The Proteobacteria accounted for 20.75-46.90% of the total phyla, while Acidobacteria was the second most abundant phylum with a range of 12.57-32.31%; these two bacteria phyla comprised more than half of the bacterial community. Moreover, the relative abundance of Proteobacteria was highest in S9 and lowest in S1, while the relative abundance of Acidobacteria was greatest in S1 and lowest in S9. Additionally, the relative abundances of Chloroflexi and Bacteroidetes were greatest at S1 and S6, respectively. Compared with the control group, the relative abundance of Acidobacteria and Chloroflexi increased, while Bacteroidetes and others decreased. The non-metric multidimensional scaling (NMDS) index is shown in Figure 2. According to the species information contained in the sample, it is reflected in the multi-dimensional space in the form of points, and the degree of difference between different samples is reflected by the distance between the points and the point, and finally obtains the spatial location map of the sample [51]. The control group and the surrounding sampling sites of the secondary lead plant had a far distance, which announced that the secondary lead smelting activity had severe effects that reduced the soil bacterial diversity in the surrounding soil.
To further investigate the composition of the bacterial community's structure, a bar plot of the taxonomic distribution of phyla in soil samples was generated (Figure 3). In the surrounding area of the secondary lead plant, Proteobacteria, Acidobacteria, Chloroflexi, Actinobacteria, Gemmatimonadetes, Nitrospirae, Bacteroidetes, Latrscibacteria, Planctomycetes, Verrucomicrobia, Saccharibacteria, Firmicutes, GAL15, and WS6 occupied more than 90% of the total sequences of each soil sample. The Proteobacteria accounted for 20.75-46.90% of the total phyla, while Acidobacteria was the second most abundant phylum with a range of 12.57-32.31%; these two bacteria phyla comprised more than half of the bacterial community. Moreover, the relative abundance of Proteobacteria was highest in S9 and lowest in S1, while the relative abundance of Acidobacteria was greatest in S1 and lowest in S9. Additionally, the relative abundances of Chloroflexi and Bacteroidetes were greatest at S1 and S6, respectively. Compared with the control group, the relative abundance of Acidobacteria and Chloroflexi increased, while Bacteroidetes and others decreased.

Soil Bacterial Community Response to Changes in Environmental Factors
The correlation between the environmental variables and soil bacterial community matrices could be compared by the Mantel test [52]. In the present study, the soil physicochemical properties, enzyme activities, and heavy metal concentrations of the 16 surrounding areas of the secondary lead plant were divided into three environmental matrices. These groups were then analyzed with the soil The non-metric multidimensional scaling (NMDS) index is shown in Figure 2. According to the species information contained in the sample, it is reflected in the multi-dimensional space in the form of points, and the degree of difference between different samples is reflected by the distance between the points and the point, and finally obtains the spatial location map of the sample [51]. The control group and the surrounding sampling sites of the secondary lead plant had a far distance, which announced that the secondary lead smelting activity had severe effects that reduced the soil bacterial diversity in the surrounding soil.
To further investigate the composition of the bacterial community's structure, a bar plot of the taxonomic distribution of phyla in soil samples was generated (Figure 3). In the surrounding area of the secondary lead plant, Proteobacteria, Acidobacteria, Chloroflexi, Actinobacteria, Gemmatimonadetes, Nitrospirae, Bacteroidetes, Latrscibacteria, Planctomycetes, Verrucomicrobia, Saccharibacteria, Firmicutes, GAL15, and WS6 occupied more than 90% of the total sequences of each soil sample. The Proteobacteria accounted for 20.75-46.90% of the total phyla, while Acidobacteria was the second most abundant phylum with a range of 12.57-32.31%; these two bacteria phyla comprised more than half of the bacterial community. Moreover, the relative abundance of Proteobacteria was highest in S9 and lowest in S1, while the relative abundance of Acidobacteria was greatest in S1 and lowest in S9. Additionally, the relative abundances of Chloroflexi and Bacteroidetes were greatest at S1 and S6, respectively. Compared with the control group, the relative abundance of Acidobacteria and Chloroflexi increased, while Bacteroidetes and others decreased.

Soil Bacterial Community Response to Changes in Environmental Factors
The correlation between the environmental variables and soil bacterial community matrices could be compared by the Mantel test [52]. In the present study, the soil physicochemical properties, enzyme activities, and heavy metal concentrations of the 16 surrounding areas of the secondary lead plant were divided into three environmental matrices. These groups were then analyzed with the soil

Soil Bacterial Community Response to Changes in Environmental Factors
The correlation between the environmental variables and soil bacterial community matrices could be compared by the Mantel test [52]. In the present study, the soil physicochemical properties, enzyme activities, and heavy metal concentrations of the 16 surrounding areas of the secondary lead plant were divided into three environmental matrices. These groups were then analyzed with the soil bacterial community matrix by the Mantel test. It showed that the soil bacterial community surrounding the secondary lead plant was related to enzyme activities (r 2 = 0.45, p = 0.001), and heavy metal contents (r 2 = 0.27, p = 0.024), but no clear correlation with soil physicochemical properties (r 2 = −0.13, p = 0.391).
Furthermore, an RDA analysis (R Development Core Team, Vienna, Austria) was conducted to verify the exact effects of physicochemical properties, heavy metals, and enzymes in soil bacterial communities (Figure 4). The results showed that soil pH (r 2 = 0.64, p = 0.001) and organic matter (r 2 = 0.58, p = 0.008) had a strong positive correlation with soil bacterial community diversity (Figure 4a). The soil enzymes protease (r 2 = 0.72, p = 0.003), hydrolase (r 2 = 0.76, p = 0.001), and dehydrogenase (r 2 = 0.62, p = 0.002) were observed to have a significant effect on soil microbial community diversity (Figure 4b). Moreover, this study found that Zn (r 2 = 0.71, p = 0.001) and Cd (r 2 = 0.43, p = 0.022) had a significant effect on the soil microbial community structure (Figure 4c). bacterial community matrix by the Mantel test. It showed that the soil bacterial community surrounding the secondary lead plant was related to enzyme activities (r 2 = 0.45, p = 0.001), and heavy metal contents (r 2 = 0.27, p = 0.024), but no clear correlation with soil physicochemical properties (r 2 = −0.13, p = 0.391). Furthermore, an RDA analysis (R Development Core Team, Vienna, Austria) was conducted to verify the exact effects of physicochemical properties, heavy metals, and enzymes in soil bacterial communities (Figure 4). The results showed that soil pH (r 2 = 0.64, p = 0.001) and organic matter (r 2 = 0.58, p = 0.008) had a strong positive correlation with soil bacterial community diversity (Figure 4a). The soil enzymes protease (r 2 = 0.72, p = 0.003), hydrolase (r 2 = 0.76, p = 0.001), and dehydrogenase (r 2 = 0.62, p = 0.002) were observed to have a significant effect on soil microbial community diversity (Figure 4b). Moreover, this study found that Zn (r 2 = 0.71, p = 0.001) and Cd (r 2 = 0.43, p = 0.022) had a significant effect on the soil microbial community structure (Figure 4c).

Potential Resistance of Soil Bacteria to Heavy Metals Contamination
According to the results of the PLI, soil pollution grades near the secondary lead plant could be divided into three groups. The low-pollution group was named L (S6 and S8). The medium-pollution group was named M (S1, S2, S3, S4, S5, S7, S10, S11, S12, S13, and S16); and the high-pollution group was named H (S14, S15, and S9). The unaffected sample was used as the control group (C). The average value of the relative abundance of each group in the genus and the phylum level is shown in Figure 5. As the concentration of heavy metals increased, the relative abundance of Sphingomonas and Nitrospira increased (Figure 5a). In the presence of higher heavy metal concentrations, Planctomycetes, Verrucomicrobia, and Latescibacterial exhibited lower relative abundances, while the relative abundance of Nitrospirae and Gemmatimonadetes increased. However, Actinobacteria, Bacteroidetes, Chloroflexi, Acidobacteria, and Proteobacteria, which dominated the bacterial community, had no apparent correlation with changing heavy metal concentrations. Actinobacteria fluctuated with slight

Potential Resistance of Soil Bacteria to Heavy Metals Contamination
According to the results of the PLI, soil pollution grades near the secondary lead plant could be divided into three groups. The low-pollution group was named L (S6 and S8). The medium-pollution group was named M (S1, S2, S3, S4, S5, S7, S10, S11, S12, S13, and S16); and the high-pollution group was named H (S14, S15, and S9). The unaffected sample was used as the control group (C). The average value of the relative abundance of each group in the genus and the phylum level is shown in Figure 5. As the concentration of heavy metals increased, the relative abundance of Sphingomonas and Nitrospira increased (Figure 5a). In the presence of higher heavy metal concentrations, Planctomycetes, Verrucomicrobia, and Latescibacterial exhibited lower relative abundances, while the relative abundance of Nitrospirae and Gemmatimonadetes increased. However, Actinobacteria, Bacteroidetes, Chloroflexi, Acidobacteria, and Proteobacteria, which dominated the bacterial community, had no apparent correlation with changing heavy metal concentrations. Actinobacteria fluctuated with slight decreases as heavy metal concentrations rose. The changes in Acidobacteria and Chloroflexi were similar, while Proteobacteria first declined and then increased (Figure 5b). Furthermore, the heatmap analysis of each heavy metal with the first 50 bacterial phyla was calculated ( Figure 6). The Pb content had negative relationships with both Bacteroidetes (r 2 = −0.58, p = 0.019) and TM6 Dependentiae (r 2 = −0.54, p = 0.033), while the decrease of the two bacteria phyla could be used to characterize the increase of Pb pollution. While Pb content and Gracilibacteria (r 2 = 0.56, p = 0.024) were positively correlated and had a close relationship, Cu and Planctomycetes (r 2 = −0.51, p = 0.048) were negatively correlated with each other, but were positively correlated with the Nitrospirae (r 2 = 0.52, p = 0.042). The Zn and GAL15 (r 2 = −0.77, p = 0.001) were negative correlated. Similarly, Cr content had a negative correlation with Fusobacteria (r 2 = −0.51, p = 0.045). However, no more than two heavy metals could be characterized by a unified group of bacteria. decreases as heavy metal concentrations rose. The changes in Acidobacteria and Chloroflexi were similar, while Proteobacteria first declined and then increased (Figure 5b). Furthermore, the heatmap analysis of each heavy metal with the first 50 bacterial phyla was calculated ( Figure 6). The Pb content had negative relationships with both Bacteroidetes (r 2 = −0.58, p = 0.019) and TM6 Dependentiae (r 2 = −0.54, p = 0.033), while the decrease of the two bacteria phyla could be used to characterize the increase of Pb pollution. While Pb content and Gracilibacteria (r 2 = 0.56, p = 0.024) were positively correlated and had a close relationship, Cu and Planctomycetes (r 2 = −0.51, p = 0.048) were negatively correlated with each other, but were positively correlated with the Nitrospirae (r 2 = 0.52, p = 0.042). The Zn and GAL15 (r 2 = −0.77, p = 0.001) were negative correlated. Similarly, Cr content had a negative correlation with Fusobacteria (r 2 = −0.51, p = 0.045). However, no more than two heavy metals could be characterized by a unified group of bacteria.  To identify the impact of heavy metal pollution on the soil bacterial community, the top 100 genera were analyzed through the Pearson correlation heatmap (Figure 7). According to the primary decreases as heavy metal concentrations rose. The changes in Acidobacteria and Chloroflexi were similar, while Proteobacteria first declined and then increased (Figure 5b). Furthermore, the heatmap analysis of each heavy metal with the first 50 bacterial phyla was calculated ( Figure 6). The Pb content had negative relationships with both Bacteroidetes (r 2 = −0.58, p = 0.019) and TM6 Dependentiae (r 2 = −0.54, p = 0.033), while the decrease of the two bacteria phyla could be used to characterize the increase of Pb pollution. While Pb content and Gracilibacteria (r 2 = 0.56, p = 0.024) were positively correlated and had a close relationship, Cu and Planctomycetes (r 2 = −0.51, p = 0.048) were negatively correlated with each other, but were positively correlated with the Nitrospirae (r 2 = 0.52, p = 0.042). The Zn and GAL15 (r 2 = −0.77, p = 0.001) were negative correlated. Similarly, Cr content had a negative correlation with Fusobacteria (r 2 = −0.51, p = 0.045). However, no more than two heavy metals could be characterized by a unified group of bacteria.  To identify the impact of heavy metal pollution on the soil bacterial community, the top 100 genera were analyzed through the Pearson correlation heatmap (Figure 7). According to the primary To identify the impact of heavy metal pollution on the soil bacterial community, the top 100 genera were analyzed through the Pearson correlation heatmap (Figure 7). According to the primary analysis, Zn content had a substantial correlation with Aquicella (r 2 = −0.60, p = 0.014), Ramlibacter (r 2 = 0.58, p = 0.018), RB41 (r 2 = −0.52, p = 0.043), SM1A02 (r 2 = −0.65, p = 0.006), and some unclassified or no-rank genus which could belong to Micromonosporaceae or GAL15. Meanwhile, Aquicella (r 2 = −0.61, p = 0.023), and SM1A02 (r 2 = −0.5, p = 0.049) were negative correlated with Cu content. However, some bacterial genera were strongly correlated with Cu, such as Nitrospira (r 2 = 0.51, p = 0.042) and Polycyclovorans (r 2 = 0.52, p = 0.025). There was also a correlation between Cd and Luedemannella (r 2 = −0.50, p = 0.05). The concentration of Cd was significantly correlated with the abundance of Steroidobacter (r 2 = 0.69, p = 0.003). Finally, it was observed that Cr exhibited a negative correlation with Luedemannella (r 2 = −0.50, p = 0.006). However, not a single genus was positively correlated with Cr. The bacteria correlating with pollution by specific heavy metals might be an alternative and effective way of forecasting the distribution of heavy metal soil pollution.

Discussion
We have found that secondary lead smelting activities might cause some changes on the soil physicochemical properties, enzyme activities, heavy metal concentrations, and bacterial communities.
Soil pH, organic matter, available potassium, nitrate nitrogen, and available phosphorus had different changes. According to the pH values, the soil was slightly alkaline, with a mean value of 7.3. It was lower than the background value, ranging from 7.8 to 8.5 [53], which means that secondary lead smelting activities led to soil acidification [54]. This might be due to the emission of sulfur dioxide in the waste smoke and the discharge of acidic waste water during the process of dismantling lead-acid batteries [4,6]. Soil organic matter in all 16 samples is also lower than the control group, which is similar to previous studies [2,4,26]. It covers all four levels of soil organic matter standards that are defined in the second soil survey of Xuzhou [53]. These very large gaps of soil organic matter in different groups also showed the complexity of soil degradation under Pb contamination [54]. At same time, there is a trend of potassium and nitrogen accumulation surrounding the secondary lead plant. Unfortunately, available phosphorus does not show regularity. The result is similar to the previous research conclusion [55,56].
Some soil enzyme activities changed near the secondary lead smelting. Soil urease and protease activity appeared to be enhanced across the whole sampling site. Three other enzymes had no similar changes. This might contribute to the close relationship between urease, protease, and nitrogen cycling [57][58][59].
Secondary lead smelting might cause serious heavy metal contamination [4,12,24,26]. In our study, results have shown that all sites were contaminated, with most sites exhibiting medium or high pollution with heavy metals. Similar to this, industrial smelting activities-such as secondary copper, zinc, and aluminum-have also led to serious heavy metal pollution [60][61][62]. In addition, heavy metal concentrations in the eastern sampling sites, S14 and S15, were higher than other sites, which might be because the eastern sampling sites were near the wastewater outlet of the secondary lead plant [4].
Furthermore, soil bacterial community diversity and composition was reduced near the secondary lead plant. Other studies also reported the decreasing trend in the similar lead smelting activities [2,3,8,21].
This study also found that pH, organic matter, protease, hydrolase, dehydrogenase, Zn, and Cd had a close relationship with the soil bacterial community. Previous studies have yielded both similar and dissimilar results. For instance, one study found that there were no obvious relationships between microbial diversity and soil physicochemical properties [52]. However, another study observed that soil microorganisms were significantly correlated with changes of physicochemical properties [63]. These conflicting results in the previously-mentioned studies might be due to different environmental factors measured in different studies and situations. Setting a pH gradient ranging from 4.0 to 8.3 in an arable soil demonstrated that soil pH has a strong positive influence on soil microbial community structure [64]. Soil organic matter significantly affected the composition and diversity of microbial communities [54]. Additionally, some studies on soil enzyme activity indicated that soil enzyme activity was the primary elements of soil microbial metabolism [65][66][67]. For example, they found that rhizobia has a close relationship with protease and dehydrogenase [66], while proteolytic bacterial communities with showed a close relationship with protease [67].
In addition, some studies reported that heavy metal pollution did not have a simple linear relationship with microbial community structure and composition, which was similar with our results [68,69]. For example, Dmitri and Begonia found that low and medium concentrations of heavy metals promoted microbial diversity, while it would have negatively correlated with microbial richness when heavy metal concentration increased to a certain extent [70]. Doelman and Haanstra claimed long-term heavy metal pollution has a negative effect on soil microbial diversity [71]. Moreover, some studies have reported the complex relationships between specific heavy metal concentrations and bacteria communities. One study on heavy metal pollution in agricultural land observed that aerobic heterotrophic bacteria had a strong positive correlation with Zn [72]. Another found that Cu has little influence on major microbial communities in long-term exposure to heavy metal pollution [73].
Other extensive experiments have shown that Cd stress affected microbial population structure and diversity [74][75][76].
Similar to previous studies in the phylum of relative abundance, Proteobacteria was also observed to be the most dominant phylum in the heavy metal contaminated soil [77,78]. This might be due to Proteobacteria's unique biodegradable metabolism and ability to adapt to a wide range of habitats [79]. Some soil bacterial communities with potential resistance to heavy metals might be used for the bio-indicator of metal-polluted soil [80]. In this study, the Ramlibacter that were observed might be utilized to predict Zn contamination, while Nitrospira and Polycyclovorans could be used to diagnose Cu contamination, and Steroidobacter could be used to forecast Cd contamination. Judging from previous studies, Rhodobacter [81], Cyanobacteria [82,83], and Steroidobacter [52] might also prove useful to monitor soil heavy metal contamination in the future [83,84]. However, the functions and structures of these newly-discovered bacterial genera need further experimentation.

Conclusions
In this study, soil physicochemical properties, enzyme activities, heavy metal concentrations, and bacterial diversity near a secondary lead plant of Xuzhou were measured. We conclude the following points: Firstly, secondary lead smelting altered the physicochemical properties and enzyme activities. Soil acidification decreased one order of magnitude, with a mean value of 7.3. Soil organic matter also showed a downward trend, while potassium and nitrogen appeared to accumulate. Soil urease and protease activity increased in samples with heavier heavy metal pollution. The concentrations of Cu, Cd, Pb, Zn, and Cr exceed the Chinese Environmental Quality Standard for Soils (GB 15618-1995). Moreover, PLI-with a mean value of 2.69-confirmed that heavy metal contamination, especially Pb and Cd, was severe near the secondary lead plant.
Secondly, Proteobacteria, Acidobacteria, Chloroflexi, Actinobacteria, Bacteroidetes, Gemmatimonadetes, Nitrospirae, Latrscibacteria, Verrucomicrobia, and Planctomycetes comprised more than 90% of the total bacterial community. The most abundant phylum in the soil was Proteobacteria, which accounted for 20.75-46.90% of the bacterial community. Furthermore, Acidobacteria was the next most abundant phylum, with a range of 12.57-32.31% of the bacterial community. Comparing with the control site, soil microbial diversity decreased, as shown by the decrease of about 50% in the Chao1 estimator, while the Shannon diversity index decreased by approximately 20%.
In addition, soil pH, organic matter, enzyme activities, and the concentration of Zn and Cd had significant correlations with soil bacterial community diversity. Also, several genera were observed to be significantly correlated with heavy metal concentrations. Ramlibacter, Nitrospira or Polycyclovorans, and Steroidobacter were strongly correlated with Pb, Cu, and Cd, respectively. These genera might act as bio-indicators for soil remediation, but the function and structure of them will need testing in future studies.
It is anticipated that the results of this research will provide new insight in the future for reclaiming soil contamination caused by secondary lead smelting.