Initial construction on the Shibahe tailings dam (lat 35°15′~35°17′ N, long 118°38′~111°39′ E) began in 1969. This copper tailings dam is a branch of the expansive Northern Copper Mine in China’s Shanxi Province. The Shibahe tailings dam is currently subdivided into 16 smaller sub-dams. The region is under the influence of a continental monsoon climate, where the average annual temperature is 13.5°C and annual precipitation is 631mm (Jia et al., 2019). Our experiments were conducted at three of the 16 three sub-dams: the S516 sub-dam, the S536 sub-dam, and the S560 sub-dam, whose representative phytoremediation years are 50, 22, and 5 (Supplementary Table S1; Jia et al., 2019).

In April 2019, we collected a total of nine I. cylindrica litter samples from each of the three sub-dams. Litter samples were then divided into two, where one was used as initial litter (D0) for high-throughput sequencing and the other was naturally air-dried to use in litter decomposition bags. The size of the nylon mesh bags was 20cm×20cm, with an aperture of 1mm×1mm. A total of 8g of I. cylindrica litter was placed into each nylon mesh bag. In May 2019, three sample plots were established in each sub-dam for which three mesh bags were added to each plot. This provided a total of 27 litter bags (Supplementary Figure S1). Litter samples were collected after being allowed to decompose for 100days (D100), 200days (D200), and 460days (D460). This provided a total of 36 litter samples. These littler samples were subsequently divided into two. One was stored (−20°C) for high-throughput sequencing, and the other was stored (4°C) to ascertain physiochemical properties and enzyme activities.

Litter bags were collected, and sediment removed, after which the fresh weight of litter (Wi) was measured. A portion of the samples was dried at 65°C to a constant weight, and water content (P) was measured. The mass residual rate of litter was then calculated (Zhang et al., 2019). The mass residual rate (%)=W_i^*(1-P)/W₀. Here, W₀ denotes the dry weight of the initial litter bag.

An elemental analyzer (vario EL/MACRO cube, Elementar Analysensysteme GmbH, Hanau, Germany) was used to measure total carbon (TC) and total nitrogen (TN) content in litter samples. Before measuring litter pH, litter water (1:20 mass/volume) suspensions were shaken for 30min (Fioretto et al., 2000). Heavy metal (Cu, Zn, Pb, and Cd) concentrations in the litter were measured using atomic absorption spectrometry (Agilent Technologies 200 Series AA, United States). Additionally, 3,5-Dinitrosalicylic acid colorimetry was used to measure litter sucrose and cellulase content; phenol-sodium hypochlorite colorimetry was used to measure urease content; potassium permanganate titration was used to measure catalase content; and the disodium phenyl phosphate colorimetric method was used to measure phosphatase content (Jia et al., 2019).

Before being filtered through sterile membranes (0.2μm pore size; Millipore, Jinteng, Tianjin, China), we washed litter samples three separate times in a sterile phosphate buffer solution (PBS: NaCl, KCl, Na₂HPO₄, and KH₂PO₄). Filtered samples were then sealed in sterile centrifuge tubes before extracting microbial DNA. Following the manufacturer’s instructions, the E.Z.N.A.^® Soil DNA Kit (Omega Bio-Tek, Norcross, GA, United States) was used to extract microbial DNA in litter samples. Additionally, we used the NanoDrop ND-1000 UV–Vis Spectrophotometer (NanoDrop Technologies, Wilmington, DE, United States) for DNA quantification. Primers ITS1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS2 (5′-GCTGCGTTCTTCATCGATGC-3′) were used as the internal transcribed spacer (ITS) gene copy numbers for all samples. Sequencing was done at Shanghai Majorbio Bio-pharm Technology (Shanghai, China), using the MiSeq platform (Illumina, Inc., United States). We submitted raw sequencing data to the National Center for Biotechnology Information Sequence Read Archive¹ under the project accession number PRJNA764552.

QIIME was used to integrate the original FASTQ format data (Caporaso et al., 2010). USEARCH (ver. 7.²) was used for the verification and removal of chimeric sequences. The operational taxonomic unit (OTU) partition threshold was established at a 97% sequence similarity of the classification results, after which it was used for fungal community diversity and relative abundance calculations. To procure species classification data that correspond to each OTU, the RDP Classifier³ was employed for the classification and analysis of each OTU sequence. Using the UNITE8.0 fungi database, the reliability threshold was determined to be 70%.

SPSS ver. 24.0 was used to analyze the physicochemical properties of litter during different decomposition stages, while Duncan’s multiple range test was used in one-way ANOVA. A Student’s t test was used to compare both fungal diversity and richness during the different decomposition stages. Moreover, we applied non-metric multidimensional scaling (nMDS) analysis to ascertain the fungal community structure based on Bray–Curtis dissimilarity, while we used analysis of similarities (ANOSIM) to ascertain intergroup differences. Additionally, we applied variance inflation factor analysis to eradicate high multicollinearity in environmental factors employing the “vegan package” in R3.5.3. We further employed Canoco 5.0 (Microcomputer Power, United States) for redundancy analysis. Finally, Gephi (an interactive platform) was used to analyze and envisage networks. Co-correlation network properties, such as the average degree, the network density, the average clustering coefficient, the network diameter, and the average path length, were calculated (Jiao et al., 2020). The higher that the average degree, the average clustering coefficient, and the network density were, the closer the network connection was assumed to be. Additionally, a lower path length and network diameter were indicative of a closed network connection (Ma et al., 2016).

The decomposition rate of I. cylindrica litter was higher during the early decomposition stage but then decreased after 100days of decomposition (Supplementary Figure S2). During the latter decomposition stage, the range of the remaining litter mass was from 27.05 to 67.14%, and the mass residual rate of litter in the S536 sub-dam was significantly higher compared to the S560 and S516 sub-dams (p<0.05).

The nutrient content of litter decreased as litter decomposition progressed (p<0.05). The C/N ratio reached a peak after 100days of decomposition and then significantly decreased. During the early decomposition stage, the litter C/N ratio in the S536 sub-dam was significantly higher compared to the other sub-dams. Heavy metal (Cu, Zn, Pb, and Cd) content gradually accumulated as decomposition progressed. and except for Zn, these heavy metals significantly increased throughout the decomposition process (p<0.05). Moreover, litter pH increased throughout the decomposition process in the S516 sub-dam (p<0.05), reaching a maximum after 200days of decomposition in the S536 and S560 sub-dams (Supplementary Table S2).

No consistent change was observed in I. cylindrica litter enzyme activities throughout the entire decomposition process in all sub-dams (Supplementary Table S3). However, catalase activity in litter significantly exceeded the initial decomposition stage (D0) in the S536 and S560 sub-dams. Litter cellulase activity in these two sub-dams was significantly higher at D100 than the other decomposition stages, while urease activity reached a maximum at D200 (Supplementary Table S3). Sucrase activity in litter significantly increased throughout the decomposition process in the S536 sub-dam, while it first increased before decreasing again throughout the decomposition process in the S516 and S560 sub-dams (p<0.05). Compared to the other sub-dams, sucrase activity in the S560 sub-dam was significantly higher (p<0.05) after 200days of decomposition (D200).

A total of 993 fungal OTUs were obtained from litter samples based on 97% sequence similarity. Specie richness (i.e., the ACE and Chao1 indices) of the litter fungi community decreased at D200 before increasing again in the S516 sub-dam (p<0.05). The Shannon, ACE, and Chao1 indices decreased after 100days of decomposition (D100) in the S536 sub-dam (p<0.05). However, the richness index of the fungi community in litter decreased at the D460 stage in the S560 dam (p<0.05; Figure 1).

Results from this study showed that common OTU numbers in each sub-dam throughout the different decomposition stages were 100 (S516), 30 (S536), and 32 (S560), accounting for 10.02, 5.23, and 6.78% of the total OTUs, indicating compositional differences in fungi communities during the different decomposition stages. We found 215 and 146 special OTUs in the litter, accounting for 37.46 and 30.93% of total OUTs at the D0 stage in the S536 and S560 sub-dams, respectively, and 76 special OTUs in the litter were found at the D0 stage in the S516 sub-dam. This was indicative of the significant differences between litter fungi communities and litter decomposition processes (Figure 2).

Ascomycota and Basidiomycota were the dominant fungi in I. cylindrica litter, with a 99.41% relative fungal community abundance. Ascomycota was the dominant fungi during the initial litter decomposition stage (D0), with an 88.68% relative abundance. At a class level, Dothideomycetes was the dominant fungal group at the D0 stage, with a 49.19% relative abundance (Figures 3, 4). Dothideomycetes abundance decreased as litter decomposition progressed (p<0.05; Figures 3, 4). On the other hand, Sordariomycetes significantly increased as litter decomposition progressed (p<0.05). Similarly, Eurotiomycetes also increased with litter decomposition and was affected by litter decomposition in the S516 sub-dam (p<0.05; Figures 3, 4). Overall, the relative abundance of Agaricomycetes peaked at D100 (Figure 3).

We observed significant differences in the composition of fungal communities in the litter at a genus level. In the S516 sub-dam, Cladosporium was the dominant fungal genus during the initial litter decomposition stage (D0), and it decreased as litter decomposition progressed (p<0.05; Figure 4E). However, Cercophora, Scytalidium, Dichotomopilus, and Knufia all increased as litter decomposition progressed. The dominant fungi genera were Talaromyces and Trichoderma during the intermediate and later stages of litter decomposition, respectively (Figure 3). Overall, we observed significant differences among Trichoderma, Cladosporium, Alternaria, and Knufia throughout the different decomposition stages (p<0.05; Figure 4H).

We constructed a genus level-based fungal community co-correlation network to explore changes in fungal community relationships during the litter decomposition process (Figure 5). Results showed that the fungal community network contained 120 nodes and 270 edges during the initial decomposition stage (D0). The number of nodes and edges increased after 100days of decomposition (D100) but decreased after 200days of decomposition (D200). Throughout the different decomposition stages, the proportions of positive fungal community correlations reached 100%, indicating that fungal communities coexisted during the litter decomposition process. Results showed that the average degree, the network density, and the average clustering coefficient of the fungal network were all highest at the D100 stage, indicating that degrees of connectivity and interaction among fungal communities were highest during the early litter decomposition stage. Additionally, to a certain extent, the degree of modularization reflected the degree of community functional diversity. The degree of fungal network modularization during the different decomposition stages was relatively high, which indicated that the functional diversity of the fungal community was high during the litter decomposition process (Table 1). The key fungal groups that maintained fungal network connectivity also changed during the litter decomposition process. Botrytis had the highest betweenness centrality (BC) value at the D0 stage. Cladostachys was gradually replaced by Clonostachys and Phaeosphaeria as litter decomposition progressed. The role that Devriesia played was important in maintaining relationships among fungal communities at the D460 stage (Table 2).

At an OUT level, nMDS analysis was conducted on fungal communities at all decomposition stages. Results showed that I. cylindrica litter decomposition significantly affected fungal community structure during the different recovery stages (S516: R²=0.623, p=0.001; S536: R²=487, p=0.003; S560: R²=0.614, p=0.001; Figure 6). Canonical correlation analysis results showed that physicochemical properties and enzyme activities in litter significantly affected the dominant fungi in litter (p<0.05), while driving factors affecting fungal communities in litter also varied during the different recovery stages (Figure 7). In the S516 sub-dam, driving factors that caused changes to the fungal community structure were Zn (R²=0.776, p=0.003), C/N (R²=0.740, p=0.006), TC (R²=0.734, p=0.004), and Pb (R²=0.662, p=0.016). Moreover, the key factors that affected fungal community structure in the S536 sub-dam were Cu (R²=0.731, p=0.001) and Zn (R²=0.579, p=0.016). Driving factors that significantly affected the fungal community structure in the S560 sub-dam were Cu (R²=0.896, p=0.002), TC (R²=0.667, p=0.01), and C/N (R²=0500, p=0.041). Overall, TN, heavy metal content (Cu, Zn, Pb, and Cd), pH, and enzyme activities (sucrose and catalase) significantly correlated with fungal community composition (p<0.05). Variance partitioning analysis results showed that litter properties and extracellular enzyme activities accounted for 13.15 and 7.10% of fungal community structure, respectively. Litter properties during the litter decomposition process primarily affected changes in fungal community structure, while the Cu content (R²=0.877, p=0.002) and pH level (R²=0.866, p=0.002) were the main ecological factors that affected fungal community structure (Supplementary Figure S3).

In this study, Trichoderma and Clitopilus positively correlated with sucrase in the S516 sub-dam (Figure 8A), which can potentially promote cellulose decomposition. Boubovia, Knufia, and Phialophora positively correlated with sucrase, and Marchandiomyces significantly correlated with cellulase in the S536 sub-dam (p<0.05; Figure 8B), indicating the important roles that these fungi groups may play in maintaining ecosystem carbon cycling. Generally, Dichotomopilus, Trichoderma, Knufia, Phialophora, Oxyporus, and Monocillium significantly and positively correlated with heavy metal content (p<0.05), indicating that these genera possess a certain heavy metal tolerance. Moreover, these fungal genera significantly and positively correlated with sucrase and catalase content (p<0.05), indicating the critical role that they play in litter decomposition within this copper tailings area (Figures 8C,D).

Litter decomposition rates closely correlate with their corresponding chemical properties, while TC, TN, total phosphorus (TP), and ecological stoichiometry in a litter affect litter decomposition rates (Sun et al., 2019). It has been reported that litter decomposition rates positively correlate with the initial TN and TP content in the litter but negatively correlate with total organic carbon, lignin, and cellulose content (Cao et al., 2016). In this study, the initial TN content in I. cylindrica litter in the S516 and S560 sub-dams is significantly higher compared to the S536 sub-dam, providing more N for microbial growth. This could potentially be the reason that litter decomposition rates in the S516 and S560 sub-dams were higher compared to the S536 sub-dam. Results show that litter decomposition strongly affected Cu, Zn, Pb, and Cd accumulation, which is consistent with a previous study (Sheng, 2009). It has been reported that multivalent metal ions can form highly stable complexes with humus during litter decomposition (Yue et al., 2019). Therefore, more metal ions bind to humus as litter decomposition progress. This indicates that litter may act as an effective metal storage pool; namely, litter absorbs metals from contaminated soil, which may be highly significant in the remediation of heavy metal contaminated soil.

Results show that fungal community diversity and richness significantly decreased as litter decomposition progressed. It has been reported the role that fungal communities play primarily occurs during the early litter decomposition stage (Baldrian, 2017). Thus, fungal richness gradually decreases as litter decomposition progresses. Dominant litter microbes are the primary decomposers in ecosystem material cycling processes, which determines litter decomposition rates (Gavito et al., 2021). Ascomycota is the dominant fungal phylum during the early litter decomposition stage, a phylum that is considered an early fungal decomposer (Zhang et al., 2017). Moreover, in this study Ascomycota is the dominant fungal phylum in all samples and is gradually replaced by Basidiomycota as litter decomposition progressed, which consistent with results from a previous study (Purahong et al., 2016). Dothideomycetes is known to secrete a variety of cellulase and hemicellulose (Wang et al., 2019a,b). Moreover, sugars, soluble starch, and cellulose in litter gradually decomposed as litter decomposition progressed, which lead to an increase in lignin accumulation that results in the colonization of Aspergillus and other microbial groups during the latter litter decomposition stage, namely microbial groups that have the capacity to decompose lignin (Bonanomi et al., 2019). Cladosporium and Alternaria are the dominant fungal genera during the early litter decomposition stage, while Trichoderma and Knufia significantly increase during the litter decomposition process. Being widely distributed throughout the soil, Cladosporium and Alternaria can utilize small-sized molecular sugars and starch. Trichoderma is a typical cellulolytic fungus with strong cellulose- and hemicellulose-decomposing abilities (Yan, 2011). In this study, Trichoderma significantly correlates with heavy metals, sucrase, and catalase, which indicate that this genus possesses a certain level of heavy metal tolerance. Additionally, Trichoderma is a highly significant fungal genus in litter decomposition in heavy metal contaminated areas.

The D100 stage yields the maximum node and edge number values and connectivity values in the fungal network, which is indicative of the complexities of interactions between the fungal communities when reaching their highest values. The nutrient content of litter is higher during the early decomposition stage, and an adequate amount of nutrients promote an increase in microbial biomass, providing greater opportunities for species to interact (Zhou et al., 2020). Results show that resource and information transmission rates are higher, and an increase in functional diversity is higher compared to a relatively simple network (Wagg et al., 2019). Additionally, complex microbial networks also improve fungal tolerance to environmental interference while being able to maintain stable microbial communities (Shang et al., 2018). Overall, litter fungi community stability is highest during the early litter decomposition stage, after which interactions among fungal communities decrease, further confirming that fungal communities may be most active during the prophase stage of litter decomposition.

Furthermore, BC values reflect how nodes connect to other nodes. Generally, the larger the BC value is, the more important role that the node plays in maintaining stable network connectivity (Chen et al., 2019). On a class level, results show that Sordariomycetes and Dothideomycetes gradually replace Leotiomycetes as litter decomposition progressed. This is consistent with the changes observed in fungal community composition. This consistency shows that the interactive ability of more abundant fungi is higher within fungal communities (Zhan et al., 2021). However, it must be noted that the key fungi within our network analysis are based on statistical analysis and therefore do not represent real-world relationships among fungal communities, which remain to be verified (Freilich et al., 2018). Therefore, future studies should selectively exclude some key species in verifying the role they play in species interactions and microbial community functions (Zheng et al., 2021).

Environmental factors (e.g., litter quality, pH, temperature, and water content) significantly affect the fungal community structure during litter decomposition processes (Chen et al., 2020a, 2021). Results from this study show that the nutrient content, the heavy metal content, and the pH level in litter significantly correlate with fungal community composition. Additionally, the Cu content and the pH level are driving factors of fungal community structure. It has been previously reported that TC and TN content in litter significantly correlates with fungal community structure (Zeng et al., 2017). Moreover, N content contributes to the synthesis of proteins and nucleotides as well as the synthesis of other macromolecules that are vital to physiological microbial functions (Purahong et al., 2015). Thus, N typically acts as a limiting factor in litter decomposition, while also affecting the microbial community structure associated with N transformation processes (Lucas-Borja et al., 2019). In soil ecosystems, pH has been reported to be a key environmental factor that affects microbial community composition and has a considerable effect on microbial community structure and functional pathways (Lucas-Borja et al., 2019). A previous study has shown that the fungal community structure and the functional pathway were both mainly driven by litter pH and TC during litter decomposition (Wang et al., 2019a,b), indicating that pH also plays an important role in litter microbial community assemblages.

Clearly, at low concentrations, heavy metals are necessary for microbial growth, and the role they play in a variety of biological metabolic processes is important (Chen et al., 2020b). However, at higher concentrations, heavy metals impact microbial composition, functional genes, and microbial diversity (Chen et al., 2018). Results from this study show that Dichotomopilus, Trichoderma, Knufia, Phialophora, Oxyporus, and Monocillium significantly and positively correlated with heavy metal, sucrase, and catalase content (p<0.05). It can therefore be concluded that soil heavy metal content may affect extracellular enzyme activities by instigating changes in fungal community composition, which will ultimately have an effect on I. cylindrica litter decomposition.