Living Fungi in an Opencast Limestone Mine: Who Are They and What Can They Do?

Opencast limestone mines or limestone quarries are considered challenging ecosystems for soil fungi as they are highly degraded land with specific conditions, including high temperature, prolonged sunlight exposure, and a lack of organic matter, moisture, and nutrients in soil. In such ecosystems, certain fungi can survive and have a crucial function in maintaining soil ecosystem functions. Unfortunately, we know very little about taxonomic diversity, potential functions, and the ecology of such fungi, especially for a limestone quarry in a tropical region. Here, we characterized and compared the living soil fungal communities in an opencast limestone mine, including mining site and its associated rehabilitation site (9 months post-rehabilitation), with the soil fungal community in a reference forest, using the amplicon sequencing of enrichment culture. Our results showed that living fungal richness in the quarry areas was significantly lower than that in the reference forest, and their community compositions were also significantly different. Living fungi in the mining sites mostly comprised of Ascomycota (Eurotiomycetes and Sordariomycetes) with strongly declined abundance or absence of Basidiomycota and Mucoromycota. After nine months of rehabilitation, certain taxa were introduced, such as Hypoxylon spp. and Phellinus noxius, though this change did not significantly differentiate fungal community composition between the mining and rehabilitation plots. The majority of fungi in these plots are classified as saprotrophs, which potentially produce all fifteen soil enzymes used as soil health indicators. Network analysis, which was analyzed to show insight into complex structures of living fungal community in the limestone quarry, showed a clear modular structure that was significantly impacted by different soil properties. Furthermore, this study suggests potential taxa that could be useful for future rehabilitation.


Introduction
Mining is considered an excessive ecosystem disturbance. Complete removal of vegetation and soil layers during mining activities remarkably changes the ecosystem in terms of both environmental condition and resources, causing loss in biodiversity and important soil elements [1][2][3]. According to Bell et al. [4], a mining site is considered the

Identification of Living Fungal Community by Amplicon Sequencing of Enrichment Culture
The samples collected from limestone quarry in this study did not provide a high quality and quantity of eDNA and the amplification of microbial gene failed [6]; thus, total fungal diversity cannot be investigated by amplicon sequencing of the eDNA. With the limitations, we enriched the soil fungal community on culture media before taxonomical identification by amplicon sequencing. The amplicon sequencing of enrichment culture method, as previously described by Sansupa et al. [17], was employed in this study. Although, it is noted that this method provided only the viable and culturable part of the Soil physicochemical properties, including soil organic matter, soil texture, macro-and micro-nutrients of the study area, were measured and presented in Sansupa et al. [6]. Briefly, forest soil was considered as clay (10% sand, 13% silt, 77% clay) with pH (H 2 O) 7.79 and contained a higher level of organic matter (6.65% SOM) and nutrients (i.e., 0.33% total N, 126 mg/kg P and 326 mg/kg K) than mine and rehabilitation plots (Figure 1b). On the other hand, soil texture in mine and rehabilitation area is considered as sandy loam (64% sand, 18% silt, 18% clay) with pH (H 2 O) 8.82 and sand clay (47% sand, 18% silt, 35% clay) with pH (H 2 O) 8.55, respectively. The mine substrate contained 0.42% SOM, 0.02% total N, < 0.05 mg/kg P and 33.20 mg/kg K, whereas rehabilitation substate contained 0.94% SOM, 0.05% total N, 3.44 mg/kg P and 72.57 mg/kg K (Figure 1b). In the forest, mine, and rehabilitation plots, soil moisture was approximately 23%, 2%, and 4%, respectively ( Figure 1b). Methods used to measure these soil properties are presented in Table S3.

Sample Collection
In each study site, 5 square plots (5 m × 5 m) were set up at a minimum distance of 20 m apart. Each square plot represents a biological replication in the study sites (n = 5). Soil samples were taken in June 2018. A total of 5 subsamples were collected to 10 cm depth, using an auger 10 cm in diameter. The subsamples were bulked into one composite sample and filtered through a sieve of 2 mm. These samples were kept in an icebox during transportation and subsequently used for the living fungal identification method within 24 h.

Identification of Living Fungal Community by Amplicon Sequencing of Enrichment Culture
The samples collected from limestone quarry in this study did not provide a high quality and quantity of eDNA and the amplification of microbial gene failed [6]; thus, total fungal diversity cannot be investigated by amplicon sequencing of the eDNA. With the limitations, we enriched the soil fungal community on culture media before taxonomical identification by amplicon sequencing. The amplicon sequencing of enrichment culture method, as previously described by Sansupa et al. [17], was employed in this study. Although, it is noted that this method provided only the viable and culturable part of the community, it was proposed as an alternative method for low microbial abundance samples and can reveal both dominant and rare taxa in the total community [6,17,20]. However, as the original method was demonstrated on bacteria, necessary modifications that included changing culture media and increasing incubation time, were made to encourage fungal growth. In detail, a living fungal community was identified in the following steps.

Enrichment of Soil Fungi
Soil fungi were enriched by three culture media including, potato dextrose agar (PDA), yeast malt agar (YM), and Dichloran Rose Bengal Chloramphenicol (DRBC) agar. In detail, 1 g of soil was added to 9 mL sterilized 0.85% NaCl and shaken thoroughly. A 100 µL of soil suspension was added to culture media and incubated at 25 • C for 7 days. Subsequently, all colonies grown on each culture media were collected and mixed in one collection tube. The colony mixtures were kept at −20 • C until further analyses.

DNA Extraction and Amplicon Sequencing
DNA was extracted from 300 µL of the colony mixture using a NucleoSpin ® Soil DNA extraction kit, according to the manufacturer's instructions. DNA samples were then amplified at internal transcribed spacer 2 (ITS2) region using forward primer ITS3F 3 -GCATCGATGAAGAACGCAGC-5 and reverse primer ITS4R 3 -TCCTCCGCTTATTGATATGC-5 [21]. Sequencing was performed using an Illumina MiSeq platform. The amplification and sequencing steps were performed at Macrogen, South Korea. Raw sequence datasets are available in the National Center for Biotechnology Information (NCBI) under BioProject accession number PRJNA753229.

Sequencing Data Analysis
Sequencing data analysis was performed using MOTHUR [22] and the Standard Operating Procedure (SOP) custom-analysis work-flow [23]. Raw sequence reads with a minimum overlap at 20 nucleotides were assembled using simple Bayesian algorithm (threshold = 0.6) as implemented in PANDAseq [24]. The assembled reads were then filtered for high-quality reads (length ≥ 200 nucleotide, Phered score ≥ 20). The chimeric sequences were detected using the UCHIME algorithm [25] and removed from the dataset. The non-chimeric datasets were clustered at 97% similarity into operational taxonomic units (OTUs) using CD-HIT-EST algorithm [26] and assigned taxonomy based on UNITE database v.7 [27] using the naive Bayesian classifier [28]. After that, rare OTUs, including singletons, doubletons, and tripletons, were removed. The datasets were normalized to 65,000 reads/sample using "rrarefy" in vegan package [29] and used for further analyses.

Functional Prediction
The potential functions associated with fungi were predicted using FungalTraits [30] and PICRUSt2 [31]. The FungalTrait database contains 10,210 fungal genera, which are classified into 17 cryptic lifestyles/traits. The fungal lifestyles/traits were assigned based on genus level. On the other hand, the PICRUSt2 predict functional potential of fungi based on marker gene sequencing profile. The gene families for ITS sequencing data were annotated corresponded to Enzyme Classification numbers (EC numbers). Specifically, this study emphasized the potential enzyme activity (based on detected gene families) of 15 soil enzymes, which could be an indicator for soil health [32], including Acid phosphatase, Alkaline phosphatase, Alpha-amylase, Alpha-N-acetylglucosaminidase, Amidase, Arylsulfatase, Beta-glucosidase, Cellulase, Chitinase, Endo-1,4-beta-xylanase, Laccase, Pectin lyase, Peroxidase, Urease and Xylan 1,4-beta-xylosidase.

Statistical Analysis and Co-Occurrence Network Analysis
All statistical analyses were performed in PAST [33] and R programming [34]. The significant differences of fungal communities in forest, mine, and rehabilitation plots were tested using nonparametric multivariate analysis of variance (NPMANOVA). Fungal community composition based on abundance (Bray-Curtis dissimilarity), and presence/absence (Jaccard dissimilarity measure) data were visulised by Non-metric multidimensional scaling (NMDS) ordinations. Subsequently, soil physicochemical parameters were fitted to the fungal community composition using "envfit" in vegan package. Goodness-of-fit (R 2 ) and significant value (p-value) were presented in Table S4. The significant parameters (p < 0.05) were included in the ordination plots. Furthermore, the functional profiles based on all predicted enzyme activity were visualized by principal component analysis (PCA). Differences in the functional profiles among the three study sites were tested by NPMANOVA. Moreover, the difference between 15 important enzyme activities in the 3 study sites were tested using ANOVA (where data were normally distributed) or Kruskal-Wallis (where data were not normally distributed). Besides, the difference between numbers of OTUs associated with each fungal trait (OTUs rich Traits) in each study site was tested by ANOVA.
Co-occurrence network analysis was performed to observe the relationships of fungal community in extreme environment (mine and rehabilitation plots). Spearman-rank correlations were calculated between all fungal OTUs detected in the extreme environment. Correlation coeffeicient (σ) more than 0.7 with a significant p-value (p < 0.05) were considered as robust and used to generate co-occurrence network. The network properties was caculated with igraph package [35] and subsequently visualized by Gephi [36] using an undirected network and Frauchterman-Reingold layout. In this network analysis, node represents OTUs, while edge/link represents the connection between OTUs. Nodes were grouped into modules when they were highly connected within their own group but much less connected outside the group [37]. A network has a modular structure when the modu-larity value of the network is greater than 0.4 [37]. In this study, modules of less than five nodes were excluded.
To identify node topologies, node connectivity within a module (Z i ) and between module (P i ) were calculated using "gateway_coeff" and "part_coeff" function in brainGraph package [38]. The node topologies were classified based on Z i and P i coefficient, into four simple catagories including, peripheral nodes (Z i ≤ 2.5, P i ≤ 0.62; nodes with a few edges), connectors (Z i ≤ 2.5, P i > 0.62; node links to several modules), module hubs (Z i > 2.5, P i ≤ 0.62; node links to many nodes in their own modules), and network hubs (Z i > 2.5, P i > 0.62; node links to many nodes both inside and outside their own modles) [39,40]. The network properties, Z i and P i score were presented in Table S5. Furthermore, the fungal community composition of all living fungi in extreme environments and that of each module were analyzed with PCA for those composition with short gradient length or CA for those composition with long gradient length. Correlation between three axis score (axis 1, axis 2, and axis 3), deriving from PCA or CA, of each community and soil physicochemical properties were calculated using Spearman-rank correlation.

General Infomation: An Overview of Sequencing Analysis
A total of 1,550,544 high-quality reads were detected in this study. Specifically, 636,195 (127,239 ± 9903, mean ± SE), 478,192 (95,638 ± 8131) and 436,157 (109,039 ± 16,034) reads were detected from forest, mining, and rehabilitation plots, respectively. After normalization and taxonomic classification, 350 fungal OTUs were obtained. Futhermore, rarefaction curves of the fungal OTUs derived from the three study sites gradually reached the saturation stage indicating that the detected OTUs were sufficient to represent the living fungal richness and community composition in this study ( Figure S1).

Living Fungal Community Composition: The Distribution of Living Fungi in Mining and Rehabilitaion Plots as Compared to Forest
Fungal richness in mining and rehabilitation plots were significantly lower than that in the forest. Whilst 82 OTUs (26 ± 4, mean ± SE) and 88 OTUs (31 ± 2, mean ± SE) were respectively found in mining and rehabilitation plots, 265 fungal OTUs (105 ± 5, mean ± SE) were found in the forest. There were three fungal phyla, including Ascomycota, Basidiomycota, and Mucoromycota, detected in this study (across three study sites), but the latter was only detected in the forest. Whilst the most abundant and frequently occurring taxa in the forest were Sordariomycetes, Eurotiomycetes, Mucoromycetes, and Dothideomycetes, those in the mining and rehabilitation plots were Eurotiomycetes and Sordariomycetes. Significant disimilarlities were found in the fungal community compositions in the three study sites. According to abundance data (based on Bray-Curtis dissimilarity measure), the composition of the living fungal community in mining and rehabilitation plots were similar. Abundance taxa in the mine were Aspergillus (59%), Fusarium (35%), Hypoxylon (1.6%), and Xylaria (1.6%), whereas those in rehabilitation plots were Aspergillus (64%), Phialemoniopsis (18%), Hypoxylon (8%), and Fusarium (2%) (Figure 2). On the other hand, the community compositions in mining and rehabilitation plots were both significantly different from that in the forest (F = 1.662, p = 0.04; Figure 2a). Fungi belonging to Trichoderma (51%), Aspergillus (23%), Nectriaceae (8%), and Absidia (7%) dominated in the forest ( Figure 2b).
In addition, we found that several taxa (45 OTUs; i.e., Curvularia affinis, Fusarium solani, Aspergillus niger, Trichoderma sp., and Hypoxylon sp.) were found in all study sites, while many taxa that were detected in the forest disappeared in mining and rehabilitation plots, for example, Trichoderma spp., Candida glabrata, and Saksenaea oblongispora. In addition, we found that several taxa (45 OTUs; i.e., Curvularia affinis, Fusarium solani, Aspergillus niger, Trichoderma sp., and Hypoxylon sp.) were found in all study sites, while many taxa that were detected in the forest disappeared in mining and rehabilitation plots, for example, Trichoderma spp., Candida glabrata, and Saksenaea oblongispora. Moreover, several taxa were only detected in the limstone quarry area (Mining and Rehabilitation), for example, Aspergillus amstelodami, Aspergillus fumigatus, Humicola phialophoroides, Curvularia clavata, and Phellinus noxius (Table S6). When it comes to fungal taxa in limestone quarry sites, it is also noticed that certain taxa found unique in mine sites, for example Aspergillus alabamensis and Nigrospora oryzae, and rehabilitation plots, for example Gibberella intricans, Hypoxylon spp., and Phellinus noxius (Table S6). However, this difference was small and did not significantly differentiate overall community composition between the two sites. The results of soil physico-chemical properties fitted to the ordinations illustrated that several parameters significantly influenced the living fungal community compositions (Figure 2a,c; Table S4). Whilst percentage of sand and pH were positively correlated to the fungal communities in mine and rehabilitaion plots, moisture, SOM, and nutrients (i.e., N, P, and Mn) were positively correlated to the community in the forest (Figure 2a,c).

Predictive Functional Profiles: Fungal Traits and Enzyme Activities Detected in Mining and Rehabilitaion Plots as Compared to Forest
Two functional prediction tools were used in this study. In total, 334 out of 350 fungal OTUs were identified into a total of 11 fungal traits, including saprotrophs (unspecified saprotroph, soil saprotroph, litter saprotroph, wood saprotroph, nectar/tap saprotroph, dung saprotroph), animal parasite, epiphyte, foliar endophyte, mycoparasite, and plant pathogen. However, there were only three main traits, including mycoparasite, saprotrophs, and plant pathogen, that contained more than five associated OTUs. On the other hand, 254 out of 350 fungal OTUs were assigned to 897 enzymes, and 15 selected soil enzymes were found in all study areas.
In the harsh conditions of the quarry sites (mining and rehabilitation plots), more than 90% of fungal OTUs were assigned to at least one trait. Whilst the most abundance traits were saprotrophs and plant pathogens, followed by animal parasite, frequently occurring traits were saprotrophs and plant pathogens, followed by mycoparasite, epiphyte and animal parasites. However, only 2 traits that contain more than 5 OTUs were saprotrophs (93 OTUs) and plant pathogens (16 OTUs). According to statistical analysis, the OTUs rich Trait (number of OTUs associated with particular function/trait) of saprotrophs and plant pathogen in mining and rehabilitation plots were not significantly different from that in the forest (Figure 3a,b). However, OTUs rich Trait of mycoparasite in the forest is significantly greater than mining and rehabilitation plots (Figure 3c). The most prevalant traits in the forest were mycoparasite, followed by saprotrophs and plant pathogens (Figure 3d,g).   In this study, mining and rehabilitation plots were considered as areas with harsh soil conditions as there were conditions of drought, heat exposure, and poor soil nutrition. Here, 130 fungal OTUs, belonging to Ascomycota and Basidiomycota were discovered. The most abundant taxa found in both mining and rehabilitation plots were Sordariomycetes and A total of 897 enzymatic genes (that associated with fungal sequences), predicted by PICRUSt2, were detected. The PCA ordination considering functional profile based on the abundace of all detected enzymes presented no significant difference among the three study areas (Figure 3e). In contrast, a difference was found between the predicted functional profiles based on presence/absence OTUs data. The functional profiles in the mining and rehabilitation plots were dissimilar from those in the forest (Figure 3h). Furthermore, 15 enzymes reported as the important indicators of soil health were selected to illustrate soil functional performance of the fungi in each study site. The results showed that all 15 enzymes were found to be possibly secreted by living fungi detected in all study sites. Specifically, we found that certain enzymes, such as Beta-glucosidase, tend to have higher activity in the harsh environments of the limestone quarry as compared to the forest. However, some enzymes, such as Alpha-N-acetylglucosaminidase and Chitinase, may have more activity in the forest than in mining and rehabilitation areas (Figure 3f,i).

Living Fungi in Mining and Rehabilitation Plots
In this study, mining and rehabilitation plots were considered as areas with harsh soil conditions as there were conditions of drought, heat exposure, and poor soil nutrition. Here, 130 fungal OTUs, belonging to Ascomycota and Basidiomycota were discovered. The most abundant taxa found in both mining and rehabilitation plots were Sordariomycetes and Eurotiomycetes. On the other hand, the most frequently occurring taxa (presence/absence data) belonged to Eurotiomycetes, Sordariomycetes, Dothideomycetes, and Agaricomycetes. Aspergillus species dominated. Fungal taxa found in these area were listed in Table 1 (and  Table S6). Here, to highlight the potential use of fungi for mine rehabilitation purpose, we listed fungal taxa found in the limestone quarry and added details on the plant-related functions based on prior studies. Furthermore, the potential strategies to survive in the harsh conditions of the limestone quarry are also presented. However, notice should be taken that the stress response information of several detected taxa in this study was not shown because there were no previous reports on such topic (Table 1).

Co-Occurrence Network and Taxonomic Distribution of Living Fungi in Mining and Rehabilitation Plots
To determine the insightful structure of fungi in the harsh environment of the limestone quarry, co-occurrence network, network properties, and critical hubs were calculated and identified. In detail, the network consisted of 126 OTUs (nodes) with 702 connections (edges) ( Table S5). The average path length and average clustering coefficients were 5.79 and 0.832, respectively. A modularity index of the network was 0.79, which suggested that the network had a modular structure. Specifically, nine modular communities were generated in this study (Figure 4). Whilst Module 0, Module 1, and Module 7 were dominated by Aspergillus species, Module 2, Module 6, and Module 8 were dominated by Hypoxylon, Penicillium, and Fusarium, respectively. Based on Z i and P i scores, there were three nodes, belonging to OTU0004 (Aspergillus flavus), OTU0014 (Hypoxylon sp.), and OTU0113 (Curvularia affinis) that were found to be connectors of this network. Other nodes were classified as peripheral nodes. Furthermore, we found that each modular community was significantly correlated with different soil properties (σ > 0.7, p < 0.05; Figure 4). Total fungal community in this area was significantly correlated to SOM, N, Ca, Fe, Mn, and percentage of sand (Figure 4). there were three nodes, belonging to OTU0004 (Aspergillus flavus), OTU0014 (Hypoxylon sp.), and OTU0113 (Curvularia affinis) that were found to be connectors of this network. Other nodes were classified as peripheral nodes. Furthermore, we found that each modular community was significantly correlated with different soil properties (σ > 0.7, p < 0.05; Figure 4). Total fungal community in this area was significantly correlated to SOM, N, Ca, Fe, Mn, and percentage of sand (Figure 4). Co-occurrence network of living fungi found in limestone quarry sites (mining and rehabilitation plots). A links/edge stands for a significant correlation based on Spearman's rank correlation (σ > 0.7, p < 0.05). Node size corresponding to degree. Node color corresponding to modularity class. The links between modular community and soil properties represent the significant correlation based on Spearman's rank correlation (σ > 0.7, p < 0.05).

Discussion
The operation of an opencast limestone mine or limestone quarry had a significant impact on edaphic parameters and microbial communities. Taxonomic distribution and community composition of soil microbe, including both bacteria and fungi, were intensively disturbed by the processes [6,8]. This study illustrated living fungi response to the mining operations and early rehabilitation process. This is a case study, which carried on an opencast limestone mine in Northern Thailand and it is noted that this study can only present the living and culturable fungal community. Here, we revealed that mining decreases living fungal diversity, compared to the reference forest. No significant difference were found on overall living fungal community composition between mining area and rehabilitation plots. However, this study highlights the taxonomic distribution of living fungi detected in the limestone quarry and discusses their living strategies, potential function, and benefit to plants, which could be promising information for future rehabilitation plans.

Living Fungi and Early Mine Rehabilitation: Community Structure, Diverity and Their Associated Function on Early Rehabilitaion Plot as Compared to Mining and Forest Sites
This study revealed that there is no respose on the live fungal community in the early stages of mine reahbilitation (<1 year). The living fungal community in rehabilitation plots was comparable to that of the mining site, but considerably different from that of forest soil. This is consistent with the trend seen in living bacterial community of the same site [6]. In this case, the living bacterial and fungal communities did not provide an immediate response to mine rehabilitation. However, it should be noted that the technique utilized was limited in that only culturable fungi were discovered. Such results are unclear when it come to the eDNA-based method. To date, there has been relatively little comparison of fungal communities in young mine restoration plots and un-restored mine sites. More research is needed to answer the question of how long the restoration time should be to dramatically alter the soil microbial community in the restoration plot relative to the unrestored mine area. These could show the effect of restoration in an early age. On the other hand, several studies have been conducted on fungal community response to restoration chronosequence. Whlist Ngugi et al. [62] demonstrated that fungal communities eventually became more similar to the reference condition over time, Kane et al. [63] found no obvious trend in fungal diversity toward the reference forest. However, both investigations found that the fungal community was mostly shaped by edapic variables [62,63]. Consistently, our research revealed that soil organic matter, pH, texture, moisture, and nutrients influenced live fungal communities, with the communities in mining regions different from those in the forest. However, a little increase in soil organic matter in the restoration plot was not enough to distinguish its fungal community from that of the un-restored mine site. Still, there are some taxa, for example, Hypoxylon spp. and Phellinus noxius, that were discovered in the restoration plot but not at the mining site. These might be caused by the introduction of sub-soil stockpile and planted trees.
Based on funtional prediction tools used in this study, saprophytic fungi were abundant in all study sites, but they were especially abundant in mining and rehabilitation plots (almost 50% of all detected traits). The result was consistent with a previous study on sand mine restoration [64]. According to Naranjo-Ortiz and Gabaldón [65], saprophytic fungi can thrive in a variety of niches and are frequently found in harsh environments. This may be explained by the abundace of saprophytic fungi in the quarry sites: since quarries are specific habitats with no plant cover, low soil nutrients and low organic substrate, other fungal lifestyles, except for saprotrophs, may not be able to survive in such conditions. In contrast, a high abundance of saprophytic fungi and mycoparasite and some epiphyte, endophyte, and plant pathogens were detected in the forest. Since the forest contained multiple habitats for various fungal species, it is not surprising that several fungal traits were detected. However, it is noticed that mycorrhizae, which are important fungi in the soil ecosystem, were not found in this study and this may be caused by the limitation of the culture method employed in this study. More diverse culture media and longer incubation periods may increase the number of fungal taxa detected, increasing the method's effectiveness [17]. Furthermore, it should be noted that functional assignment in this study may not cover all possible functions contributed by the community. Previous studies have demonstrated that fungal functions, predicted by PICRUSt2 and FungalTrait, may be limited by the number of genes available in the database and poor taxonomic classification [30,31,66].

Living Fungi in Harsh Condition of the Limestone Quarry: There Are Many Way to Survive
The extreme condition in this study relates to drought, heat exposure, low nutrient availability, and slightly alkaline (~pH 8) soil, which were all caused by mining operation. We demonstrated that Ascomycota fungi, mainly Dothideomycetes and Eurotiomycetes, are often found in this area. This could be attributed to the fact that Ascomycota is a widespread and ubiquitous phylum that represents the most diffuse group of fungi and has the greatest number of currently recognized species [67]. However, our finding follows a similar pattern to earlier research that found fungi belonging to these two classes predominating in dry [68] and alkaline [69] environments. Despite the fact that this study cannot determine whether the identified taxa were active or dormant in such conditions, it is believed that detected taxa were able to survive under drought and low nutrient soil due to the utilized method. We demonstrated that fungi belonging to the Aspergillus species were the most abundant taxa found in the severe conditions of the opencast limestone mine. This corresponds with previous study that found Aspergillus DNA in a variety of severe environments, such as high salinity and cold [43]. The presence of Aspergillus species in severe environments may be explained by the availability of conidia, which are asexual spores produced by Ascomycota under unfavorable conditions. It has been demonstrated that fungal conidia can withstand a variety of environmental stressors, such as drought, high temperatures, and ultraviolet (UV) irradiation [68]. These may be induced by the accumulation of compatible solutes (i.e., manitol or trehalose) in fungal conidia, which function as a cell defender under stress [69,70]. However, Wyatt et al. [68] stated that Ascospore, an Ascomycota sexual spore, is more resistant to stress than conidia. Therefore, it could imply that one of Ascomycota fungi's strategies for survival under environmental stress is through their spores. This might be one of the reasons why 90% of the fungi detected in our study were Ascomycota. However, there may be additional strategies that can help the fungi cope with environmental stress, such as the expression of a stress response gene or the formation of thick cell walls [44,71].

Co-Occorence Network of Living Fungi in Harsh Condition of The Limestone Quarry: The Modular Community and Their Correlation with Soil Physiochemical Properties
According to the Zi and Pi values, no OTUs in this network were theoretically identified as hubs. However, the three connectors (OTU0004: Aspergillus flavus; OTU0014: Hypoxylon sp.; OTU0113: Curvularia affinis) with strong linkages to each module were discovered and might be defined as a key population of this network [72]. Furthermore, the network can be sub-structured into several modular communities. Previous studies showed that the microbial modular pattern was driven by soil physicochemical properties [73][74][75]. This was supported by our results. We demonstrated that different soil characteristics were significantly correlated with each modular community. This might suggest that members in each module have different niches [75]. Here, we showed that K, Mg, and S were all negatively correlated with Modules 0, 1, and 6, respectively. On the other hand, Module 8 was negatively correlated with SOM, Fe, and N. However, that certain modules (Module 2, Module 4, Module 5, and Module 7) did not have a significant correlation with observed soil characteristics may imply that members of these modules may exist in flexible soil conditions. However, more study is needed to clarify this issue, particularly studies that only captured active fungus.

Posible Interaction of Fungi and Plant: Opportunity for Rehabilitation Procedure
Most of the restoration on the quarry area has been done by covering the floor with topsoil [64,76,77]. The adding of topsoil could facilitate plant survival and act as microbial inoculant at the early stage of mine rehabilitation. This could be the priority procedure to consider when working on mine rehabilitation. However, in some cases, topsoil has not been a choice. Inoculation of plants with fungi that can both tolerate the extreme mine conditions and can promote their growth should be considered. We presented that several fungi surviving in the harsh environments of the limestone quarry were previously reported as facilitators to enhance plant-stress tolerance. For example, Aspergillus aculeatus and Aspergillus fumigatus were shown to aid plants in drought environments and to further promote their growth [42,45]. This shows a promising application of fungi detected in a mine area to accelerate plant growth and survival in a mine. However, more studies have to be done on the effects of fungal inoculation on each tree species.

Overall Living Fungal Community Assessment: Limitation and Recommendation for Future Work
Living fungal community in this study was investigated by amplicon sequencing of enrichment cultures. This method is beneficial for low-biomass samples where an eDNAbased method cannot be employed, as well as for rapid screening of the living microbial community and its predictive functions [17]. Although this study successfully evaluated the fungal community in each study site, it should be noted that we can only disclose culturable taxa growing in the applied media. It is recommended to apply a variety of culture media in future work to obtain more diverse taxa [17]. The identification of fungal taxa in this study based on short-read sequencing (Illumina Miseq Platform) can potentially identify the microbes deep to the genus level and, in some case, to the species level [78]. To more accurately identify the species of fungi, it is suggested to apply long-read sequencing, which was recently proposed as a promising tool for fungal species identification [79,80]. Furthermore, the application of both the traditional culture method and high-throughput sequencing would be an alternative option for future studies as it could contribute more comprehensive and better resolution in species identification and microbial distribution patterns [78,81].

Conclusions
This study reveals a living fungal community in an opencast limestone mine in Northern Thailand. The study was held in 2018, where the rehabilitation was processed for approximately 9 months. Here, we found that the rehabilitation of the opencast limestone mine, by dumping sub-soil stockpile and planting framework tree species, did not have an early (< 1 year) impact on living fungal community composition. In the limestone quarry, most of the detected fungi were saprotrophs. Ascomycota, especially Aspergillus, was one of the most abundant taxa that survived in the harsh conditions of this area. It is also suggesting that Aspergillus species could be of particular interest for future mine rehabilitation due to their ability to survive in harsh conditions and to promote plant growth. However, the plant growth information presented in this study was based on literature; further research is required to test such ability on taxa isolated from this quarry.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/jof8100987/s1, Figure S1: Rarefaction curve of observed fungal OTUs across three study sites., Figure S2: Co-occurrence network of fungal OTUs detected in harsh environment of mining and rehabilitation sites, highlighting (a) taxonomic distribution at genus level and (b) functional distribution based on FungalTrait., Table S1: List of tree species found in the reference forest., Table S2: List of planted tree species in rehabilitation sites., Table S3: Method used for soil physicochemical measurement., Table S4: Goodness-of-fit (R 2 ) and significant value (p-Value) of soil parameters fitted to NMDS ordinations of living fungal community, Table S5: Topology and node properties of cooccurrence network of fungal OTUs detected in degraded environment of mining and rehabilitation plots. Table S6: Living fungal taxa found in forest, mining, and rehabilitation plots.