The responses of soil bacterial communities and enzyme activities to the edaphic properties of coal mining areas in Central China

Soil physicochemical properties, bacterial communities and enzyme activities change with land subsidence resulting from coal mining. However, research on the responses of bacterial communities and enzyme activities to the soil properties in different degree of subsidence areas is limited. As such, we collected soil samples from a control area (C area), a moderate mining subsidence area (M area) and a severe mining subsidence area (S area) in Central China. Soil properties, such as the pH, total nitrogen (TN) content, total phosphorus (TP) content, available phosphorus (AP) content, organic matter (OM) content, and soil enzyme (urease, invertase, catalase and alkaline phosphatase) activities were measured in each sampling area at depths of 0–20 cm, 20–40 cm, and 40–60 cm. The results indicated that the soil physiochemical properties, soil urease activity, soil alkaline phosphatase activity and soil bacterial richness and diversity in the topsoil (0–20 cm) of the mining subsidence area were significantly lower than those in the C area. However, the soil enzyme activities within the deepest layer of the subsidence area were significantly greater than those of the C area. The bacterial communities within the depth of 0–20 cm were dominated by RB41, Pseudomonas, MND1, Nitrospira, Trichococcus, Sphingomonas and Dongia, whereas RB41 and Pseudomonas were the dominant species in the C area and subsidence area, respectively. Using correlation analysis, we found that the soil pH value, soil AP content and activities of the four enzymes were the main factors affecting the soil bacterial community structure. In addition, the soil nutrient contents, enzyme activities and bacterial richness and evenness decreased with increasing subsidence degree (classified by geological hazards, groundwater and landscape damage degree of coal mining subsidence). These results provide a reliable basis for environmental management of mining areas.


Introduction
Large-scale exploitation of coal resources has promoted economic development in China. However, this process has inevitably led to serious environmental problems [1]. More than 95% of coal mining occurs underground in China because the coal is buried at great depths. As a result, many ecological problems, such as land subsidence and cracking have occurred. Moreover, the soil hydraulic properties [2], soil nutrient distribution [3], and biological community [4] have been inevitably altered. Shallow hills surround the Pingdingshan mining area, and the ecological system is prone to severe soil erosion. Land subsidence aggravates soil erosion, decreases soil fertility, and leads to ecological degradation, and soil nutrients can infiltrate into deep layers in mining subsidence areas [4]. However, the vertical distribution of soil physiochemical properties in land subsidence areas has not been studied recently. As dominant regulators of soil OM decomposition and nutrient cycling, bacteria and enzymes play important roles in soil ecosystems. Enzyme activity can be used as an indicator of microbial activity [5] and plays a vital role in the decomposition of organic compounds, resulting in nutrient increases on the Loess Plateau [6]. Though the soil bacterial community and enzyme activity are sensitive to environmental change, their responses to soil properties in mining subsidence areas have not yet been studied.
The process of coal mining can alter the soil environment, such as the soil density, moisture [2] and nutrient distribution, all of which are closely related to bacterial diversity and enzyme activity. Soil enzymes are mainly derived from the activities of soil plants, animals and microbes. They are involved in various biochemical processes of soil, such as decomposition and synthesis of humus, decomposition of plant and microbial residues, hydrolysis and transformation of synthetic organic compounds, oxidation and reduction of inorganic compounds. Enzyme activities in collapsed ground were found to be lower than those of control groups whose ground was intact [7]. [8] found that the bacterial abundance was greater, and the bacterial community structure more complex in reclamation sites than in coal-excavated subsidence sites in mining fields in East China. [4] investigated the effect of post mining subsidence on soil bacterial diversity and its response to edaphic variables in the western China aeolian sand area. They found that soil microbial richness and diversity decreased due to land subsidence, and the main factors driving the changes in the bacterial community structure were the soil electronic conductivity, soil water content and soil depth. However, research rarely has been conducted on the effects of land subsidence on soil bacterial diversity and enzyme activity in hilly areas. In this study, we investigated the effects of different levels of mining subsidence on soil bacterial diversity and enzyme activities and how soil chemical properties responded in a hilly area.

Sampling site description
This study was conducted in the No.1mine, Pingdingshan city, Henan Province, Central China (113˚11 0 45@-113˚22 0 30@ E, 33˚40 0 15@-33˚48 0 45@ N) (Fig 1). These areas were investigated under the implementation of "Mine Geological Environment Protection and Rehabilitation Program" from Tian An coal mine of Pingdingshan. This region is located between Pingding and Luoniao Mountains, where the altitude ranges from 150 to 300 meters. The soil is mainly albic yellow cinnamon soil according to the soil classification system of the Food and Agriculture Organization-United Nations Educational, Scientific and Cultural Organization [9]; the soil erodes easily and contained inadequate nutrients to support vegetation. This region has a semiarid continental monsoon climate. The mean annual temperature is 14.9˚C, the mean annual precipitation is 742.6 mm, the main period of rainfall occurs from July to September, and the mean annual evaporation is 2,157.8 mm. This study area was divided into three categories according to different subsidence degrees: S area, M area and C area (no subsidence) (Fig 1).
Permits obtained for the work. The mining area is located in the suburb of the city, it has been abandoned and left unattended. Therefore, no permits were required for sampling.
Ethics statement. The study was approved by Pingdingshan Coal Industry No.1 mine. The field studies did not involve endangered or protected species.

Sample collection
The subsidence area was mined in December 2011. Seven years after subsidence (June 2018), soil samples from 0-20 cm, 20-40 cm and 40-60 cm depths were collected from the S (S1, S2, and S3, respectively) area, M (M1, M2, and M3, respectively) area and C (C1, C2, and C3, respectively) area. Before sampling, three sampling plots (2 m×2 m) were established at each site. In each plot, three composite samples as replicates were randomly collected at each soil depth. After roots and stones were carefully removed, each 500 g soil sample was divided into three parts: one was air dried for determining the soil chemical properties, one was stored at -80˚C for 16S rRNA gene analysis, and the rest was stored at 4˚C for measuring the soil enzyme activity.

Soil analyses
2.3.1 Physiochemical analysis. Soil property analysis was performed using standard soil test procedures. The soil pH values were measured at 1:2.5 soil/water ratios (w/v) via a pH meter [10]. The soil TN concentrations were determined by further grinding a subsample to particles 0.15 mm in size and igniting them in a high-induction N analyzer furnace (Elementar Americas Inc., Mt. Laurel, NJ). The TP of the soil was digested with perchloric acid and then measured by colorimetry. The soil OM was analyzed by volumetric determination with potassium dichromate, and the soil AP concentrations were measured via the molybdenum antimony colorimetric method [11].
2.3.2 Enzyme activity. The soil catalase activity was measured with the methods of [12], while the soil alkaline phosphatase activity was measured according to the p-nitrophenyl phosphate method, as described by [13]. The soil urease activity was assayed on the basis of the amount of ammonia released: after compost samples and urea were incubated together; 1 N KCl and 0.01 N HCl were mixed to extract the ammonium, which was detected via the colorimetric method [14]. The soil invertase activity was estimated by the 3, 5-dinitrosalicylic acid (DNSA) method with sucrose used as a substrate.

Statistical analysis
The QIIME 1.9.1 data analysis package was used for the 16S rRNA gene analysis. The forward and reverse reads were joined and assigned to samples on basis of their barcodes and were truncated by removing the barcode and primer sequences. Quality filtering of the joined sequences was performed, and sequences that did not fulfill the following criteria were discarded: sequence length <200 bp, no ambiguous bases, and mean quality score > = 20. The sequences were then compared with a reference database (RDP Gold database). The UCHIME algorithm was used to detect chimeric sequences, which were subsequently removed. The effective sequences were used in the final analysis. The sequences were grouped into operational taxonomic units (OTUs) by the clustering program VSEARCH (1.9.6) against the Silva 132 database that was preclustered at a 97% sequence identity. The ribosomal database program (RDP) classifier was used to assign taxonomic categories to all OTUs at a confidence threshold of 0.8. The RDP classifier uses the Silva 132 database which has taxonomic categories predicted at the species level. The sequences were flatten prior to the alpha and beta diversity statistical calculations. That is, we will take the same number of sequence from each sample and then analyze it later. The alpha diversity indices were calculated by QIIME from soil samples, the Shannon index was used to determine the bacteria diversity and the Chao1 index was used to determine the richness. Weighted and unweighted UniFrac distances were calculated to estimate beta diversity with non-metric multidimensional scale (NMDS). An unweighted pair group method with arithmetic mean (UPGMA) tree constructed on the basis of the beta diversity distance matrix.
Correlations between the soil properties and bacterial communities were determined by SPSS 22.0. Canonical correspondence analysis (CCA) was used to analyze the relationships between the bacterial community and environmental factors in R 3.5.1 in conjunction with chi-square test. Differences between the samples were calculated via one-way analysis of variance (ANOVA) in conjunction with a p test at the level of 0.05.

Soil enzyme activities
The soil enzyme activities varied with the degree of land subsidence. The soil urease activity was greatest in the M area at a depth of 40-60 cm (Fig 2A); as the degree of subsidence increased, the soil alkaline phosphatase activity decreased. The soil alkaline phosphatase activity at a depth of 0-40 cm was significantly greater than that at a depth of 40-60 cm in the C area. The opposite results were recorded in the M and S areas ( Fig 2B). The soil invertase activity at a depth of 40-60 cm was significantly greater than that at a depth of 0-20 cm in both the M and S areas ( Fig 2C). Moreover, the soil catalase activity was great throughout the topsoil in three areas, and increased with increased subsidence degree (Fig 2D).

Soil physiochemical properties
The three analyzed groups differed with respect to the soil nutrient indicators, such as the soil OM, TN, TP and AP contents and pH value ( Table 1). The soil TP and TN contents at the same depth in the C area were generally greater than those in the M and S areas; In addition, the contents were significantly greater at the 0-20 cm than at other depths in all three areas. The soil pH at the same depth was significantly greater in the C area than in the M and S areas. Moreover, the pH at 0-20 cm was significantly greater than those at 20-60 cm in the M and S areas. The soil OM content decreased significantly with increasing subsidence. However, the differences were not significant among all soil layers, indicating that the vertical distribution of OM in the soil was relatively stable. In the C area, the soil AP content was significantly highest at 40-60 cm, while the opposite occurred in the M and S areas.

Soil bacterial community composition
Proteobacteria was the dominant phylum in all samples in the different areas (22.25%-38.68%), followed by Acidobacteria (16.07%-49.57%), Actinobacteria (5.5%-12.31%) ( Fig 3A) and Rokubacteria (2.1%-13.83%). These phyla accounted for more than 70% of the bacteria within the soil samples. Proteobacteria was the most abundant phylum at the 20-40 cm depth in each sample. The relative abundance of Acidobacteria was greatest in the deep layer (40-60 cm) in the C area, and lowest in the topsoil (0-20 cm) in the M and S areas, indicating that coal mining subsidence altered the distribution of the bacterial species in the soil in the vertical direction. The differences in the relative abundance of the Actinobacteria phylum among all soil samples were not significant. In comparison, the relative abundance of the Rokubacteria phylum was greater in the M area than that in the C and S areas.
The seven most abundant genera that have been classified included RB41 (1.01%-30.28%), Pseudomonas (3.36%-12.54%), MND1 (1.96%-4.88%), Nitrospira (0.98%-5.8%), Trichococcus (1.24%-4.85%), Sphingomonas (1.16%-7.72%) and Dongia (1.08%-3.21%) (Fig 3B). The relative abundance of RB41 within the same layer was significantly greater in the C area than in the M and S areas, while that of Pseudomonas was the opposite. There were no significant differences in the relative abundance of MND1 among any of the samples from the three areas. The relative abundances of Nitrospira, Trichococcus and Dongia were greatest in the topsoil (0-20 cm) in the C area, and greatest in the deepest layer (40-60 cm) in the S area. However, there were no

Soil bacterial richness and diversity
A total of 401,972 high-quality reads were obtained from the soils in the C, M and S areas after 509,567 raw reads were filtered (ranging from 46,865 to 72,159) via the clustering program Vsearch 1.9.6, which resulted in 1,294, 1,284, and 1,296 OTUs, respectively, the clustering threshold is 97%. The sample numbers satisfied the OTU determination based on rarefaction curves (S1  Table 2). These four indices changed similarly with depth. However, in the S area, the Ace and Chao indices in the deepest layer (40-60 cm) were greater than those in the other layers. The values of the Ace, Chao, Shannon and Simpson indices were greater in the S area than in the C area at the same depth. These results indicated that the bacterial richness and diversity decreased with depth in the M area, but increased with depth in the S area.

Soil microbial community structure
The NMDS analysis (stress value <0.033) was based on the beta diversity distance matrix calculated from the species composition among all the samples (S4 Fig). The results showed that the C1, C2 and C3, M1, M2 and M3, and S1, S2 and S3 sample were grouped close together, suggesting that the soil bacterial diversity was affected mainly by the degree of subsidence. The soil bacterial diversity in each of the C and M areas was more similar than that in the S area; moreover, the distance between the deepest layer (40-60 cm) and the other layers (0-40 cm) was greater in the C and M areas than in the S area. This finding indicated that severe subsidence disrupted the soil bacterial community. Moreover, the soil bacterial community in the topsoil (0-20 cm) in the S area was significantly changed by land subsidence. A heatmap was constructed to display the relative abundances and dissimilarity in the microbial community composition among the samples at the gene level (Fig 4). For instance, the relative abundance of RB41 (30.28%) in the C3 area increased by 121.0 2%, 166.31% and 321.14% respectively and significantly compared with C1, C2 and S1 areas, whereas was less abundant in the all the depths of M area. In comparison, MIZ17 was abundant in the M areas but deficient in other areas; the relative abundance of Pseudomonas in the S2 area (9.35%) was 4.59 times, 1.61times and 1.69 times higher than that of S3, C1 and C3 areas respectively. The relative abundance of Trichococcus and Dongia were more in the S area than others.

Correlations between bacterial communities and soil properties
Pearson's correlation coefficients between soil properties and bacterial communities are shown in Table 3. RB41 was positively correlated with the pH (p<0.05) and catalase activities (p<0.05) but negatively correlated with the urease and alkaline phosphatase activities (p<0.05). Pseudomonas was significantly negatively correlated with OM (p<0.01), TP contents (p<0.05) and the urease and alkaline phosphatase activities (p<0.05) but significantly positively correlated with the catalase activity (p<0.01) and invertase activity (p<0.05), and MND1 was significantly positively correlated with AP and TP contents (p<0.05). Nitrospira was significantly negatively correlated with catalase activity (p<0.05) and invertase activity (p<0.05).
Trichococcus was significantly negatively correlated with TP and OM contents (p<0.05) but significantly positively correlated with the catalase activity (p<0.01) and invertase activity (p<0.05). There was a significant positive correlation between Sphingomonas and AP content (p<0.01). Dongia was significantly negatively correlated with the catalase activity (p<0.05). All bacteria were significant correlated with the urease and alkaline phosphatase activities (p<0.05). Therefore, the soil chemical properties and enzyme activities in the study area had significant effects on the bacterial communities. In contrast, The TP, AP, and OM contents  The objective of the CCA was to identify the main environmental factors affecting the bacterial community structure (Fig 5). Among the soil chemical properties, the soil pH and AP and TP contents and the soil catalase and urease activities played vital roles in forming the bacterial community structure (S3 Table). RB41 was significantly positively correlated with the soil pH and soil TN, TP and OM contents in the C area. Moreover, there was a significant correlation between Sphingomonas and the soil AP content in the M area. Trichococcus, Pseudomonas and MND1 were significantly correlated with the activities of four enzyme in the S area (Fig 5). The variance results from the principal component analysis showed that the simulation model had a good degree of fit (p<0.001, S1 Table), and the significant eigenvalues of the variance analysis are shown in S2 Table. The bacterial communities were divided into four groups, which indicated that the degree of subsidence and soil depth affected the soil bacterial community structure.

Effects of land subsidence on soil enzyme activities and physiochemical properties
In this study, with increase of soil depth, the soil invertase, urease and alkaline phosphatase activities decreased significantly (Fig 2). These findings were in accordance with previous results [7]. This phenomenon can be explained by the distribution of soil nutrients within different layers, for the enzyme reactions are based on nutrients [15]. Land subsidence decreased the soil urease and alkaline phosphatase activities but increased the soil catalase and invertase Table 3. Pearson's correlation coefficients between the relative abundance of the 7 dominant bacterial genera and soil properties.  activities (Fig 2). Soil urease and alkaline activities are the main factors that regulate N and P cycling [16]. N and P losses caused by subsidence caused a decrease in enzyme activities [4]. Invertase catalyzes the hydrolysis of sucrose into glucose and fructose to provide carbon for plants and microbes. One possible explanation might be that the soil OM content in the C area was greater than S area, and soil microbes may synthesize additional catalase and invertase to adapt to unfavorable environments. Land subsidence severely deformed the surface and redistributed the nutrients in the vertical direction of soil. In the present study, with increase of soil depth and subsidence extent, the soil OM, TN and TP contents decreased (Table 1). These results were consistent with those from the study of [3]. Land subsidence and cracking can alter the soil bulk density and porosity [17; 18]. With the precipitation soluble nutrients such as C, N and P are lost along many pores, which leads to the OM, TN and TP decrease in the subsidence areas. In addition, subsidence can aggravate soil erosion and OM composition [19]. In the present study, soil pH in the M and S areas increased with increase of soil depth. But the soil AP content decreased in the M and S areas and increased in the C area ( Table 1). The soil AP content in the C area was affected mainly by the runoff infiltration. However, a large amount of AP was released onto the surface from the decomposition of phosphoric minerals in the acidic environment in the M and S area.

Variation in soil bacterial community composition by land subsidence
Bacterial community structure are affected by many ecological factors. For instance, climate conditions can affect the abundance of genes from Actinobacteria, Planctomycetes and Proteobacteria [20]. A sequencing study revealed that Actinobacteria can outcompete other dominant groups, such as Acidobacteria under arid conditions [21]. [22] reported that different plant species change the composition of soil bacterial community by affecting the soil pH. Proteobacteria, Acidobacteria, Actinobacteria and Rokubacteria were the dominant bacterial phyla in the present study, accounting for approximately 80-85% of the total OTUs. This result implies that these phyla may play important roles in coal mining areas. The climate and environmental conditions were identical among the sites in this study, but the subsidence levels differed; thus, the conditions were appropriate for analyzing how soil properties affecting soil bacterial diversity and bacterial communities.
The process of coal mining can alter the soil environment, through actions such as acidifying, compacting and eroding the soil, which directly affected the bacterial community. The results of the present study showed that the soil bacterial community composition varied among the three studied areas. The diversity and richness of the bacterial community decreased with increasing of subsidence level and soil depth. Though the numbers of bacterial communities were lowest in the S area, they play an important role in the regulation of biogeochemical cycles [23; 24; 25]. In addition, anthropogenic soil disturbances (such as agriculture, mining, and the disposal of residues or waste) shift ecosystem equilibria, forcing microorganisms to adapt to new habitats [26]. In the present study, the dominant bacterial communities comprised RB41, Pseudomonas, MND1, Nitrospira, Trichococcus, Sphingomonas and Dongia in the C and S areas. RB41, belongs to Acidobacteria and exhibited a relatively range of tolerance to pH (up to 10). Some studies have reported that RB41 might play an important role in maintaining the metabolic and biogeochemical function of soil under long-term low-nutrient stress conditions. Our results were the opposite; one explanation is that the pH was greater in the C area than that in the S area, and the conditions were more suitable for RB41 to survive in the C area. However, Pseudomonas was the most abundant bacteria in the S area, and members of this genus have demonstrated a great deal of metabolic diversity. The higher abundance of Pseudomonas is linked with soil health [27; 28].The organisms in the Pseudomonas genus are mostly free-living bacteria that are widely distributed among soil, water, and marine environments and are present on the skin of animals, including humans. They are strict aerobes and move via one or more polar flagella. Land subsidence results in the creation of many soil pores, satisfying the bacterial demand for oxygen.

Response of soil bacterial community and enzyme activities to soil properties
In the present study, the effect of soil urease, catalase, and invertase activities on the bacterial community structure was greater than that of alkaline phosphatase activity (Fig 5). Soil enzyme activities are the main way for microbes to decompose nutrients and OM [29] and are relatively more representative of bacterial physiology [30; 31].
Physiochemical soil parameters are the major influencing factors for soil microbial diversity [32; 33]. We found that various physicochemical soil factors had different effects on the bacterial diversity in areas affected by coal mine subsidence; specifically, soil pH and AP content were most significant in shaping soil bacterial communities. Many studies have demonstrated that pH is a major factor determining soil bacterial community composition [34; 20]. pH was a key factor affecting the diversity, structure, interaction, and function of rhizosphere bacterial communities in crop soil under control conditions [35]. The soil pH affects the protein synthesis of Acidobacteria, Actinobacteria and Cyanobacteria [20]. In addition, Acidobacteria had the function of degrading plant reidue multimers [36; 37], which provided nutrient input to those that survive under acidic conditions. Some study have also found that soil pH was correlated with soil bacterial phylogenetic diversity at the continental scale, with peak diversity in soil with near-neutral pH [38]. That may be the major reasons why soil pH was closely correlated with soil bacterial communities.
In addition, the soil AP content was positively correlated with the abundance of MND1 and Sphingomonas (belong to Proteobacteria). The growth rate hypothesis emphasizes that the growth of most bacteria is related to the need for P for ribosomal RNA synthesis, prompting the bacterial community to enter the soil [39]. Bacteria play important roles in the mineralization of organic P and dissolution of inorganic P. Microbial inorganic P solution occurs primarily through release of organic anions, which causes precipitated P-metal complexes to solubilize, releasing inorganic P [40; 41]. Bacteria capable of this mode of solution include members of Pseudomonadaceae, Rhizobiaceae, Caulobacteraceau, Comamonadaceae, Burkholderiaceae, and Enterobacteriaceae within Proteobacteria [42].
This study found land subsidence influenced soil bacterial community richness and diversity, these effects were realized by changes of soil physiochemical properties and enzyme activities. However, the correlations between the soil properties and the microbial communities differ among soil systems. For instance, [43] reported a significant correlation between soil Actinobacteria and Proteobacteria and soil OM, TN and TP. [10] observed the effects of soil properties on the bacterial communities present in soil from the Shengli Oilfield and found that soil TP was positively correlated with Actinobacteria but negatively correlated with Firmicutes.

Conclusion
Mining-induced subsidence significantly decreased the soil enzyme activities in the topsoil, especially those of soil invertase, urease and alkaline phosphatase, In addition, the soil pH, and soil TN, TP and OM contents decreased in response to subsidence, and the soil properties and bacterial community diversity and richness at a depth of 0-20 cm were more susceptible to coal mining. The dominant bacterial communities comprised RB41, Pseudomonas, MND1, Nitrospira, Trichococcus, Sphingomonas and Dongia. The main environmental factors determining the soil bacterial community in the coal mining areas of Central China were the pH and AP content, which determine the protein and RNA synthesis of bacteria. In short, the soil nutrients, enzyme activities and soil microbial community structure were disturbed by land subsidence. Therefore, it is necessary to take measures to improve the environment of subsidence areas.