Changes in Soil Organic Carbon Fractions and Fungal Communities, Subsequent to Different Management Practices in Moso Bamboo Plantations

Moso bamboo (Phyllostachys pubescens) has an extremely fast growth rate and major carbon sequestration potential. However, little information is available on the dynamics of soil C accumulation and fungi communities related to different management practices. Here, we investigated changes in the soil organic carbon (SOC) fractions and fungal communities of a Moso bamboo plantation under three different management practices (M0: undisturbed; M1: extensively managed; and M2: intensively managed). Compared with M0, SOC levels were reduced by 41.2% and 71.5% in M1 and M2, respectively; furthermore, four SOC fractions (C1: very labile; C2: labile; C3: less labile; and C4: nonlabile) and the carbon management index (CMI) were also significantly reduced by plantation management. These practices further altered fungal communities, for example, by increasing Basidiomycota and Mortierellomycota, and by decreasing Ascomycota and Rozellomycota. Pyrenochaeta, Mortierella, Saitozyma, and Cladophialophora were identified as keystone taxa. Soil fungal communities were significantly related to the pH, NH4-N, AP, C3, and the C4 fractions of SOC. Random forest modeling identified soil C3 and Mortierella as the most important predictors of the CMI. Our results suggest that reducing human interference would be beneficial for fungal community improvement and C sequestration in Moso bamboo plantations.


Introduction
Bamboo is an important forest resource in subtropical and tropical regions [1] and accounts for 0.80% of the global forest area [2]. Approximately 641.16 million hectares of bamboo forests exist in China, 72.96% of which consist of Moso bamboo [3]. Moso bamboo usually achieves its growth (10-20 m height) and diameter (breast height 8-16 cm) within 35-40 days after its shoots emerge from the soil in the spring [4,5]. Because of its special growth characteristics, this bamboo species is an important non-wood forest product in China [6] and has major carbon (C) sequestration potential [7,8]. Yen and Lee [8] found that Moso bamboo forests have a higher C sequestration rate (8.13 ± 2.15 Mg ha −1 year −1 ) than Chinese fir forests (3.35 ± 2.02 Mg ha −1 year −1 ). To obtain higher economic benefits, management practices have been implemented in the bamboo plantations [9,10].
Soil represents the largest terrestrial C pool, offering higher C storage than the sum of atmospheric and vegetation C storage [11]. Small changes in the soil C pool have a substantial impact on the global C balance [12]. Forest ecosystems account for more than half of the soil organic carbon (SOC) in terrestrial ecosystems [13,14], and are regarded as the most promising reservoirs for storing additional C and reducing the CO 2 load [15]. SOC pools comprise labile and non-labile C pools [16,17]. Of these, the labile C pool is subject to rapid turnover rates that influence nutrient cycling, soil quality, and productivity [18]. Conversely, the non-labile C pool is slowly altered by soil microbes [19] and is important for C sequestration [20]. As part of the labile C pool, oxidizable organic carbon (OC) could be an indicator of early dynamic changes to SOC [20,21], because it is more sensitive to changes in the physicochemical properties of soil than total organic carbon (TOC) [22].
Soil fungi are primarily involved in the decomposition of plant material, especially the lignocellulosic components, which are relatively recalcitrant to bacteria. Typically, a succession of fungal species occurs on new substrates when they enter soil, based on the initial decomposition of simpler compounds, followed by the subsequent degradation of more complex polymers [23]. Fungi play a crucial role in SOC dynamics [24,25]. For example, they produce oxidative enzymes that decompose the recalcitrant biopolymers of plant litter [25,26]. Clemmensen et al. [27] found that fungal community shifts might affect soil C decomposition and sequestration, which are closely related to other soil parameters.
Previous studies have shown that management practices can affect SOC dynamics in Moso bamboo plantations [10,28,29]. However, to our knowledge, few studies have been conducted on the relationships between oxidizable OC fractions and soil fungal communities. In this study, we compared the effects of different management strategies on the SOC fractions and fungal communities in Moso bamboo plantations. Our specific objectives were as follows: (1) to determine the effects of different plantation management strategies on SOC fractions and, consequently, on soil C sequestration ability; (2) to evaluate the shifts in the fungal communities of Moso bamboo under different plantation management strategies and the relation thereof to SOC fractions and other soil factors.

Experimental Site and Sample Collection
The field trial was conducted in Tianhuangping town, Anji County, Zhejiang Province, China. The meteorological data of this study area were shown in Yang et al. [10]. The experimental area has a subtropical monsoon climate, with an annual average temperature and sunshine of 17 • C and 1943 h, respectively. This region has 224-240 frost-free days, at altitudes of 300-380 m. The soil is classified as a Ferralsols [30].
Moso bamboo plantations subjected to the following three different management strategies were selected for evaluation: unmanaged (M0), extensively managed (M1), and intensively managed (M2). More information about these bamboo plantations was described by Yang et al. [10]. Briefly, the M0 plantation had not experienced any human management for over 60 years, and gradually developed into a mixed forest. Conversely, certain practices-including the selective and regular harvesting of bamboo trunks and shoots, every two years-were maintained in the M1 plantation, whereas the M2 plantation was subjected to annual selective harvesting, understory vegetation removal, and fertilization, during mid-late June.
In May 2021, we established three comparable stands (M0, M1, and M2) with similar initial site conditions for sample collection. Six 20 × 20 m plots were demarcated in each selected bamboo stand. Within each plot, five S-shaped soil profiles [31] were collected and carefully mixed to form a composite sample for that plot, after being passed through a 2 mm sieve. The fresh soil samples were then divided into two parts, one of which was stored at −80 • C for microbial analysis. The other was air-dried for physicochemical property analysis.

Calculation of Associations among Parameters
The carbon management index (CMI) was calculated using the following method, described by Blair et al. [35]: Microbial DNA was extracted using the Fast DNA ® Spin Kit for soil (MP Biomedicals, Santa Ana, CA, USA). The ITS1 region was amplified using the primers of ITS1F (5 -CTTGGTCATTTAGAGGAAGTAA-3 ) and ITS2 as a reverse primer (5 -GCTGCGTTCTTCATCGATGC-3 ). The PCR amplification, library construction, and Illumina MiSeq sequencing were performed by Shanghai Majorbio Bio-pharm Technology Co., Ltd. (Shanghai, China), following the standard protocols [36,37].

Statistical Analyses
All statistical analyses were conducted in R (v4.1.0; http://www.r-project.org/, accessed on 18 May 2021). One-way analysis of variance (ANOVA), followed by least significant difference (LSD) were performed to calculate the difference among the three sites in soil properties, microbial alpha diversity indices, and relative abundances of fungal-dominated phyla and genera. Rarefaction techniques provide a way to adjust the differences in library sizes across samples and aid comparative analysis of alpha diversity [44]. All samples were rarefied to the lowest number of sequences (30,807). Calculation of alpha diversity indices (Shannon and Chao1) and hierarchical clustering were performed in the 'Vegan' [45] package. A circos plot was drawn using the 'circlize' [46] package, and the network was generated and visualized using 'ggClusterNet' [47] package. Only robust correlations (Spearman's correlation coefficient > 0.6 and p-value < 0.05) were in the network. The nodes with high degree in the network were identified as keystone species [48]. Redundancy analysis (RDA) was carried out in the 'Vegan' [45] package, to assess the effects of soil factors on the fungal communities. The soil parameters used in the RDA analysis were preselected by the 'bioenv' function, and significant soil parameters were determined using the 'envfit' function in the vegan package. Spearman's correlation test and the Mantel test were performed using the 'ggcor' package [49], to evaluate the relationships between soil factors and fungal communities. Random forest modelling [50] was applied to identify the main drivers of CMI by using the 'randomForest' [51] package, whereas the significance of the model and the predictor importance were determined using 'therfutilities' [52] and 'rfpermute' [53] packages. Linear regression was employed to assess the correlation between the CMI and its important predictors using the 'basicTrendline' [54] package.

Soil Properties
Plantation management practices had a significant effect on all evaluated soil properties (Table 1). Compared with samples from the M0 and M2 plots, the pH was highest (p ≤ 0.05) in the soil from the M1 plots, in the following order: M1 > M0 > M2. The M0 soil had higher (p ≤ 0.05) SOC, NH 4 -N, and NO 3 -N contents than that from the other two sites, and the SOC and NH 4 -N in the M1 soil were significantly (p ≤ 0.05) higher than in the M2 soil. Compared with M0 and M1, the soil from the M2 plots had significantly (p ≤ 0.05) increased AP and AK, and there was a significant (p ≤ 0.05) difference in AK between the M0 and M1 soil.
A comparison of the dominant phyla and genera is presented in Figure 1b and Table  S1. M1 and M2 had reduced (p ≤ 0.05) the relative abundances of Ascomycota, Rozellomycota, Archaeorhizomyces, and Russula, relative to M0. The M1 and M2 soil had a higher (p ≤ 0.05) relative abundance of Basidiomycota, Mortierellomycota, Trichoderma, and Mortierella than the M0 soil. Compared to M0 and M1, Metarhizium was significantly (p ≤ 0.05) lower and Saitozyma was significantly (p ≤ 0.05) higher in soil from the M2 plots (Table S1). We used two indices (the observed number of OTUs and the Shannon index) to assess the fungal alpha diversity (Figure 2a). A significant (p ≤ 0.05) difference was observed in the number of OTUs among the three groups in the following order: M0 > M1 > M2 (Figure 2a). Compared to M0, the Shannon index increased (p ≤ 0.05) in M1 and decreased (p ≤ 0.05) in M2 (Figure 2a). A hierarchical cluster analysis, based on Bray-Curtis distances, showed that the soil samples from the M0, M1, and M2 sites were clustered and differentiated into three significant groups (Figure 2b). The ANOSIM test also revealed significant (p ≤ 0.05) differences between M0, M1, and M2.
A comparison of the dominant phyla and genera is presented in Figure 1b and Table S1. M1 and M2 had reduced (p ≤ 0.05) the relative abundances of Ascomycota, Rozellomycota, Archaeorhizomyces, and Russula, relative to M0. The M1 and M2 soil had a higher (p ≤ 0.05) relative abundance of Basidiomycota, Mortierellomycota, Trichoderma, and Mortierella than the M0 soil. Compared to M0 and M1, Metarhizium was significantly (p ≤ 0.05) lower and Saitozyma was significantly (p ≤ 0.05) higher in soil from the M2 plots (Table S1).
To further investigate the fungal co-occurrence patterns within each soil preparation, three networks were constructed based on the OTU level ( Figure 4). The network properties of the soil fungal communities are summarized in Table 3. The co-occurrence network analysis showed that the number of nodes and edges in the M2 and M1 network were higher than in the M0 networks; M1 and M2 networks had the same number of nodes, but M2 had an increased number of edges, compared to the M1. These indicated that the M2 network was comparatively more complex, relative to the M0 and M1 networks. Additionally, the M2 network had a higher average degree and average clustering coefficient than the M0 and M1 networks ( Table 3). The number of negative links in the M0, M1, and M2 networks were 63, 111, and 74, respectively, accounting for 40.65%, 46.06%, and 12.03%, respectively, of the total number of corresponding links ( Figure 4, Table 3). Edge width corresponds to the Mantel's R value and the edge color denotes the statistical significance.
To further investigate the fungal co-occurrence patterns within each soil preparation, three networks were constructed based on the OTU level ( Figure 4). The network properties of the soil fungal communities are summarized in Table 3. The co-occurrence network analysis showed that the number of nodes and edges in the M2 and M1 network were higher than in the M0 networks; M1 and M2 networks had the same number of nodes, but M2 had an increased number of edges, compared to the M1. These indicated that the M2 network was comparatively more complex, relative to the M0 and M1 networks. Additionally, the M2 network had a higher average degree and average clustering coefficient than the M0 and M1 networks ( Table 3). The number of negative links in the M0, M1, and M2 networks were 63, 111, and 74, respectively, accounting for 40.65%, 46.06%, and 12.03%, respectively, of the total number of corresponding links (Figure 4, Table 3).

Effects of Plantation Management Practices on Soil Properties and SOC Fractions
This study showed that SOC, C1, C2, C3, and C4 were reduced in M2 soil, compared with that from the M0 and M1 plots. Fontaine et al. [55] suggested that plant litter and root exudates are the principal sources of soil C. Another study showed that increasing plant diversity could significantly increase the amount of root exudates [56]. We also found that M1 significantly increased the soil pH, compared with M0. Girkin et al. [57] suggested that high concentrations of root exudates could cause a greater reduction in soil pH. Our previous study showed that bamboo plantations subjected to M0 experienced less disturbance than those subjected to M1 and M2, resulting in increased plant diversity and plant residue input [10]. Thus, changes in understory vegetation and litter inputs could explain the changes in the soil pH, as well as in the SOC and its fractions. Moreover, M2 soil had the lowest pH, mainly because of the use of N fertilizers, which are the primary drivers of soil acidification [58,59]. The soil from the M1 and M2 plots had lower NH4-N and NO3-N concentrations than the M0 soil. This may be because NH4-N and NO3-N are easily absorbed by plants [60] and plantation management increases the N uptake of Moso bamboo [61]. In contrast, the AP and AK contents increased significantly under M2, which was partially associated with the use of P and K fertilizers in these plantations.  Table S2). ** p ≤ 0.01; * p ≤ 0.05; ns p > 0.05. IncMSE-the increase in mean square error.

Effects of Plantation Management Practices on Soil Properties and SOC Fractions
This study showed that SOC, C1, C2, C3, and C4 were reduced in M2 soil, compared with that from the M0 and M1 plots. Fontaine et al. [55] suggested that plant litter and root exudates are the principal sources of soil C. Another study showed that increasing plant diversity could significantly increase the amount of root exudates [56]. We also found that M1 significantly increased the soil pH, compared with M0. Girkin et al. [57] suggested that high concentrations of root exudates could cause a greater reduction in soil pH. Our previous study showed that bamboo plantations subjected to M0 experienced less disturbance than those subjected to M1 and M2, resulting in increased plant diversity and plant residue input [10]. Thus, changes in understory vegetation and litter inputs could explain the changes in the soil pH, as well as in the SOC and its fractions. Moreover, M2 soil had the lowest pH, mainly because of the use of N fertilizers, which are the primary drivers of soil acidification [58,59]. The soil from the M1 and M2 plots had lower NH 4 -N and NO 3 -N concentrations than the M0 soil. This may be because NH 4 -N and NO 3 -N are easily absorbed by plants [60] and plantation management increases the N uptake of Moso bamboo [61]. In contrast, the AP and AK contents increased significantly under M2, which was partially associated with the use of P and K fertilizers in these plantations.

Effects of Management Practices on Soil Fungal Community
Soil microbial diversity had a strong link to soil quality and the nutrient cycling rate [62]. Owing to their rapid response to soil changes, soil pH and nutrient levels are widely used to assess soil quality and microbial communities [63,64]. In our study, the fungal richness and diversity were lowest in the M2 soil, which may be due to the decreased soil physicochemical properties, compared with that of soil from the M0 and M1 plots. Soil microbial communities play a key role in soil function and ecosystem sustainability [65,66]. Chaer et al. [67] suggested that more abundant biodiversity indicates greater soil stability and Maček et al. [68] showed that lower fungal biodiversity causes unsustainable crop production and ecosystem instability. These studies suggested that M2 alters the stability of bamboo plantations and threatens ecosystem function. We also found that the selected soil physicochemical properties significantly influenced the soil fungal communities and, thus, were able to infer that changes in the soil properties caused by different bamboo management approaches, would contribute to variations in fungal communities. Soil pH is a key factor in shaping the soil fungal community because it may modify nutrient availability or limit fungal growth, affecting the diversity and composition of the fungal community [69,70]. In this study, the Shannon index was highest for M1 soil samples and lowest for M2 soil samples, which was consistent with their pH order, indicating that soil pH was the most influential factor in the Shannon index.
Our results demonstrated that Ascomycota and Basidiomycota were the dominant phyla in bamboo soils, which is consistent with the results of our previous study [71]. Numerous studies have reported that members of Ascomycota and Basidiomycota are key fungal decomposers in soil ecosystems [72,73]. The Basidiomycetes species can degrade lignocellulosic organic matter more effectively than other fungal groups [74,75], whereas Ascomycetes taxa have a limited ability to degrade litter containing recalcitrant lignin [76]. These findings indicated the ubiquity and importance of Ascomycota and Basidiomycetes in bamboo ecosystems, especially concerning the C cycle. Our results also indicated that M1-and M2-managed soils had a lower abundance of Ascomycota and Rozellomycota and a higher abundance of Basidiomycota and Mortierellomycota, compared to the M0 soil. Wu et al. [77] revealed that Ascomycota prefers to live in high-N soils, whereas Basidiomycota and Mortierellomycota are more suited to growth in low-N soils. According to Tian et al. [78], soil NO 3 -N content may affect the distribution of Rozellomycota. In our study, the shifts observed in these four fungal phyla were mainly due to changes in the soil N levels (NO 4 -N and NO 3 -N).
A network analysis identified Pyrenochaeta, Mortierella, Saitozyma, and Cladophialophora as keystone taxa. Previous studies have reported that the genus, Pyrenochaeta, is involved in the degradation of polysaccharides, such as cellulose, hemicellulose, pectin, and xylan [79,80]. Furthermore, Li et al. [81] confirmed that Mortierella species can compose recalcitrant substances and contribute to soil organic matter (SOM) storage. Saitozyma has the potential to decompose dead plant biomass because of its ability to incorporate C from cellulose [82]. Harsonowati et al. [83] found that Cladophialophora could promote nutrient uptake (N and K) from the soil and improve plant growth and fitness. Based on our results, keystones were the main contributors to variations in soil C content.
The results also indicated that M2 produced greater complexity of, and lower competition among the fungal network, compared with M0 and M1. Huang et al. [84] showed that the exogenous addition of glucose could decrease the complexity of fungal networks. Dundore-Arias et al. [85] found that an increased input of soil exogenous C could enhance microbial competition for C sources because of their niche overlap. The differences in plant residue input caused by the three plantation management approaches, evaluated herein, may partly contribute to shifts in the soil fungal network. Moreover, 87.97% of the fungal OTUs in the M2 soil were positive connections, revealing that the network was unstable [86]. This phenomenon may be due to the fungal OTUs producing co-fluctuations and positive feedback with environmental changes [87,88].

Effects of Plantation Management Practices on Soil CMI
The CMI reflects dynamic changes in soil C and has been used to assess the effect of plantation management practices on soil quality [89,90]. A higher CMI value suggests that the soil has higher potential to store C and reduce losses [90,91]. Moreover, higher CMI values reflect that soil C is being rehabilitated, whereas lower values indicate that the system is declining [35]. Our results indicated that M2 resulted in the lowest CMI among the three plantation management practices, indicating that soil managed in this manner had a low-quality C pool and could be at risk of degradation. A random forest analysis revealed that soil C3 and Mortierella were the best predictor variables of the CMI. Studies have indicated that C3 has stable molecular structures and changes slowly through microorganism activity, thereby playing an important role in SOC storage [18,92]. Mortierella is involved in the decomposition of plant litter and aromatic hydrocarbons [93,94] and Li et al. [81] elucidated that the Mortierella species can produce antibacterial compounds and chemically recalcitrant melanin to delay soil C turnover, contributing to SOC storage. Therefore, C3 and Mortierella can be used as indicators for evaluating SOC storage.

Conclusions
This study showed that Moso bamboo plantations subjected to M2 plantation management practices had lower SOC, oxidizable OC fractions, and CMI values, indicating that this approach is not beneficial for the accumulation and stability of SOC. The M1 approach increased soil pH, relative to M0 and M2. We also identified keystone fungal species in the studied bamboo soils, revealing complex interrelationships among the soil physicochemical properties, oxidizable OC fractions, and shifts in the fungal communities. Moreover, soil C3 and the relative abundance of Mortierella were the most important predictors of the CMI. These findings improve our understanding of the alteration of soil oxidizable OC fractions in Moso bamboo plantations, subjected to different management practices. In conclusion, fewer soil disturbances could increase the accumulation and stability of SOC and facilitate the sustainable development of bamboo ecosystems. These results can help us design suitable management practices to improve C sequestration and, thus, mitigate climate change in forest ecosystems.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jof8060640/s1. Figure S1: Venn diagrams of soil fungal OTUs. M0, M1, and M2 indicated undisturbed, extensively managed, and intensively managed bamboo stands, respectively; Figure S2: Co-occurrence networks of fungal communities at OTU level colored by phylum in the Moso bamboo plantations under different management practices; Figure S3: Relationships between carbon management index and its important predictors. CMI, Carbon Management Index; C3, less labile fraction of oxidizable carbon; Table S1: Relative abundances of dominant genera in Moso bamboo soils under different management practices; Table S2: Keystones identified in the soil microbial communities.

Data Availability Statement:
The data that supports the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.