Changes in soil fungal community composition and functional groups during the succession of Alpine grassland

This study explores trends in soil fungal patterns, potential functions and their responses to carbon and nitrogen applications over alpine grassland succession in the Qilian Mountain area of China. The soil fungal community was characterized via Illumina sequencing of ITS genes. The FUNGuild database was used to predict functional groups in alpine grassland succession from swamp meadow to alpine meadow and steppe meadow. The levels of soil carbon and nitrogen, vegetation carbon and nitrogen, and soil enzyme activity were also assessed. Soil fungal operational taxonomic units increased from swamp meadow to alpine meadow stage and then reached a relatively stable state in steppe meadow successional stage. This result is consistent with the changing trend of sobs and Chao1 index. Moreover, the soil fungal community differed significantly between different succession stages. During succession, while most phyla of soil fungi followed linear decreasing trends, Ascomycota was the dominant fungal phylum. Its abundance increased significantly, from 60.00% in swamp meadow to 72.26% in steppe meadow. The relative abundances of Pathotroph and Saprotroph fungal functions increased with successional stages, while Symbiotroph did not change significantly. Soil ammonium nitrogen and organic carbon levels dominated the effect on the soil fungal microbial community and functional groups. These findings indicate that the fungal community shifted dramatically in the first stage(swamp to alpine meadow) of succession but then reached a relatively stable state in the second(alpine to steppe meadow) successional stage. The soil fungal function group did not follow the same successional trajectory. Soil ammonium nitrogen and soil organic carbon levels imposed the strongest influence on the fungal community and its function, respectively.


Introduction
Alpine grassland resources provide globally important ecological and social service functions by storing carbon and supporting biodiversity (Yao et al. 2019). Impacted by climate warming, alpine grasslands experience approximately twice the global average level of warming, which induces a succession of vegetation types and associated changes in soil functions . It is also generally accepted that the succession of the alpine grassland on the Qinghai-Tibet Plateau in China tends towards a drier ecosystem upon the thaw of permafrost, developing from alpine swamp meadow to meadow, steppe, and desert (Qin and Ding 2009). Changes in plants during this succession can lead to variations in soil and microbial community (Shao et al. 2017). As a centrally important part of the alpine grassland ecosystem, microbial communities play a key role during succession (Yang et al. 2019). However, microorganisms, the main contributors to the process of succession, have been rarely studied. Understanding changes in microbial composition and potential functionality can deepen knowledge of the mechanisms driving successional dynamics .
Ecological succession is closely related to the structural and functional dynamics of terrestrial ecosystems (O'Donnell et al. 2018). During ecological succession, plant communities change over time, which may also affect carbon and nitrogen cycling (Shao et al. 2017;Smith et al. 2015;Zhang et al. 2022). Previous studies have reported that the alpine grassland succession in the Qinghai-Tibet Plateau tends towards degradation. Not only plant biomass, soil organic carbon (SOC), and total nitrogen (TN) decreased, but the soil water content (SWC) and the carbon and nitrogen content of plants also changed. These effects may weaken the C sink effect of vegetation and in turn aggravate climate warming (Alhassan et al. 2018;Jiang et al. 2020;Liu et al. 2018). Many studies found that different soil microbial functional groups (or guilds) and soil enzyme activities commonly infiltrate different environments during succession. These environments are characteristic of the biochemical cycle of the investigated regions (Crowther et al. 2019;Wang et al. 2021). These findings suggest that ecosystem succession is accompanied by an outflow of carbon and nitrogen. Ecological succession not only impacts carbon and nitrogen cycle, but also influences the soil microbial community and its function. Among different successional ecologies, microorganisms are particularly susceptible, and consequently, they received much attention over the last two decades (Heo et al. 2019). For example, in forest ecosystems, the abundances of both bacteria and fungi increase over the early stages of succession but then reach a relatively stable state at later successional stages (Jiang et al. 2021). Guo et al. (2018) reported that grassland succession caused by climate warming affects the differentiation of microbial community succession; therefore, successional trajectories of microbial communities are difficult to predict under future climate warming scenarios. Additionally, they also reported that the soil physicochemical and microbiological properties shift during succession in Mediterranean humid mountain ecosystems (Company et al. 2022). Li et al. reported that in semi-arid grassland ecosystems, microbial abundance is higher at the beginning of succession . These findings suggest that the succession of different ecologies induces shifts in vegetation, soil, and microbial community, thus initiating a series of natural or secondary successions. However, how fungal communities change with the natural succession in the alpine grasslands of the Qinghai-Tibet Plateau in China has not been explored to date. To the best of our knowledge, while microbial diversity and its ecological function during succession have been reported widely, most studies focus on changes in bacterial rather than fungal communities . As an important group of the soil microbiome, soil fungal communities are more responsive to habitat changes than bacterial communities (Jiang et al. 2021). Compared with fungi, bacteria can metabolize a wider range of compounds, which may explain their relative stability compared with that of fungi (Hartmann et al. 2014). However, because of the biotrophic relationship between plants and fungi, such a change of microorganisms can effectively reflect the succession of the plant community (Zhong et al. 2018).
In addition, while fungal diversity and its ecological function over the process of succession have been widely reported, most studies focused on secondary succession induced by stresses imposed by external environmental factors (e.g., water and soil carbon). Moreover, the period of most secondary successional studies is rather short (Lee et al. 2020;Wang et al. 2021). Vegetation shows large changes in composition and production during secondary succession, which can lead to a shift in the microbial community within a short time (Holtkamp et al. 2011). However, compared with secondary succession, natural succession is a better predictor of the longterm development of ecology succession. While natural 1 3 Vol.: (0123456789) succession effectively changes the physicochemical properties of soil, it is time-consuming (Kardol et al. 2010;Sun et al. 2016). Furthermore, most studies focus on changes of structure and composition of microbial communities during succession in mixed ecosystems. For example, long temporal scales have been explored across the three stages of grass-shrub-arbor succession in terrestrial ecosystems . However, few studies focused on microbial communities in a single ecosystem on a small scale. Exploring the interaction relationship between microorganisms and environmental factors at such small spatial scale helps to address basic scientific questions of different ecologies (Feng et al. 2019). Also, understanding the main drivers of observed microbial differences across different ecosystems is an important prerequisite for designing effective management and conservation strategies (Zhao et al. 2018). Therefore, the relationships between microbial communities and environmental factors during long-term succession must be explored further. The alpine grassland of the Qinghai-Tibet Plateau is an ideal study area for such research.
The current study was conducted in the Qilian Mountain region Ecological Research base Alpine Grassland, China. This study area represents natural succession from swamp meadow (SWM) to alpine meadow (AIM) to steppe meadow (STM) and has a good sequence of undisturbed natural succession. Thus, it offers a unique landscape to examine patterns of fungal community succession. In this study, the patterns of fungal community succession were explored in three stages. These patterns were assessed using ITS microbial rRNA gene sequencing to disclose how both the structure and composition of the microbial community developed along the succession chronosequence. The objectives were to (i) evaluate the patterns of change in carbon and nitrogen levels, soil fungal communities, and fungal functional groups over this succession; (ii) determine if changes between carbon and nitrogen, as well as microbial communities are congruous; and (iii) determine the main carbon and nitrogen factors affecting fungal communities.

Study area
The study was conducted at the Qilian Mountain region Ecological Research base Alpine Grassland, located in a typical alpine meadow, Qilian County, Mule town, China (37°57′ 36′′ N, 100°19′48′′ E; 3487 m above sea level). This region has a typical plateau continental climate, and the annual average temperature is -1.7 °C. The annual average rainfall is about 614.8 mm, mainly falling during July and September . Various types of natural grassland can be found, making the area suitable for grazing. The vegetation mainly contains alpine meadow (accounting for 97% of all available grassland), followed by steppe meadow (accounting for 6.96%), and swamp meadow, which only occupies a small proportion (Cui et al. 2022). In swamp and alpine meadows, dominant plant species are Carex moorcroftii and Kobresia humilis, respectively. In recent years, because of the impact of climate warming and human encroachment, the natural grassland in this reserve has been degraded and cultivated, and degraded alpine meadows were reseeded (Fig. 1).

Experimental design
In this experiment, the 'space-for-time substitution' method was used. The three habitats SWM, AIM, and STM in the transition zone were selected to represent the three stages of the typical alpine grassland successional process in this area (Jiang et al. 2021). In September 2019, four field plots were selected per successional stage to account for treatment effects. The distance between each field plot was about 200 m. A 20 m × 20 m square was drawn in each field plot, taking the center of the square as sampling center, and three samples equidistant from the central sampling point were collected for plant investigation. All aboveground plant parts were collected in each quadrant as aboveground biomass, which were oven-dried at 65 °C for 48 h and the contents of carbon and nitrogen were analyzed . Soil samples were collected from each quadrat: eight random soil samples were collected from the 0-15 cm soil layer in each replicate field plot using a soil-drilling sampler (3.5 cm inner diameter) and were merged into one replicate sample; in total, four soil samples were obtained for each vegetation type. All soil samples were passed through a 2 mm sieve to remove other materials. Soil samples were immediately sent back to the laboratory in a cooler. Soil samples were divided into two parts to determine the chemical properties and microorganisms present in the soil. One part of the soil sample was naturally dried in the shade and then sieved through a 1 mm mesh to analyze soil chemical characteristics, while the other part was stored at -80 °C for high-throughput gene sequencing.

Carbon and nitrogen properties of plants and soil
The carbon and nitrogen contents in both soil and vegetation were analyzed within one month of collecting the samples. The contents of soil organic carbon (SOC) and total nitrogen (TN) were determined by a C and N analyzer (Elementar, Langenselbold, Germany). Soil available nitrogen was extracted with 1 M KCl, and filtrates were analyzed for soil ammonium nitrogen (NH 4 N), soil nitrate-nitrogen (NO 3 N) by a colorimetric method analyzer (CleverChem200+, Germany) (Cai et al. 2017). Chloroform fumigation extraction was used to estimate soil microbial biomass carbon and nitrogen (Yin et al. 2019). Soil samples were retrieved and dried in the laboratory. When fully dry, samples were sieved through a 1 mm sieve, urease activity (URE) was determined by the sodium phenol-sodium hypochlorite colorimetric method, and sucrase activity was determined by the 3,5-dinitrosalicylic acid colorimetric method (Liang et al. 1997). The carbon content of plants was determined by titration. Mixed plant grinding samples were disboiled by potassium dichromate oxidation and external heating. The nitrogen of plants needs to be boiled using the sulfuric acid-hydrogen peroxide method, and its content was determined by CleverChem200+ (Germany).

Bioinformatics analyses
The composition and diversity of fungi in the soil microbial community were determined using highthroughput gene detection techniques. Soil samples stored at -80 °C were transported on dry ice to Guangzhou Genedenovo Biological Technology for analysis (Illumine 2500 250 PE, USA). For each sample, total genomic DNA was extracted from 0.5 g of soil using a HiPure Soil DNA Mini Kit (Magen, Guangzhou, China). A NanoDrop 2000 (Thermo Fisher, USA) was used to assess the DNA quality. ITS rRNA was detected in the ITS2 region, and the primer sequences were KYO2F (GAT GAA GAA CGY AGY RAA ) and ITS4R (TCC TCC GCT TAT TGA TAT G). PCRs were performed in a triplicate 50 µL mixture containing 5 µL of 10 × KOD buffer, 5 µL of 2 mM dNTPs, 3 µL of 25 mM MgSO 4 , 1.5 µL of each primer (10 µM), 1 µL of KOD polymerase, and 100 ng of template DNA (Pruesse et al. 2007). Purified amplification products were mixed in equal amounts, a sequencing connector was added to construct sequencing libraries, and the Illumina PE250 library was sequenced. To ensure the reliability and validity of data, FASTP was used to filter reads from the original dataset generated by the Illumina MiSeq platform. FLASH was used to splice double-ended reads into tags and filter low-quality tags to obtain clean tags (Magoc and Salzberg 2011;. Based on the UPARSE pipeline (version 9.2.64), high-quality sequences were grouped into operational taxonomic units (OTU) with 97% similarity (Edgar 2013). The sequencing data were compared with the UNITE (ITS) database to obtain biotaxonomic information. DNA sequences involved in this study were submitted to the NCBI Sequence Read Archive (SRA) database under the BioProject number PRJNA881640. Fungal α-and β-diversity indices, including Chao1, Simpson, and Shannon indices, were calculated at the OTU level. Redundancy analysis (RDA) was conducted at the phylum level.

Statistical analyses
The Kolmogorov-Smirnov test was used to test normality and the assumptions were satisfied. The significance of differences of microbial diversity index, soil and vegetation carbon, nitrogen content, and soil enzymatic activity were analyzed using multiple comparisons (LSD method) after ANOVA, assuming significance at P < 0.05. In addition, the "ADONIS" function of the vegan package in R (999 permutations) was also used to test the significance of the separation between successional stages. ANOSIM analysis was used to detect differences within and between groups. The linear discriminant analysis (LDA) effect size (LEfSe) method was used to assess high-dimensional microbial taxa and identify differentially vegetation types according to microbial taxonomy. The steps of LEfSe analysis are as follows: the Kruskal-Wallis rank-sum test (a commonly used test method for multiple groups) was conducted for all groups. Then, the selected differential species between two groups were compared by the Wilcoxon rank-sum test (a commonly used test method for two groups). Finally, differences were selected using LDA, and the results were sorted by mapping to identify evolutionary branches. The lowest corrected Akaike information criterion values were used to select the best regression model, and the variance inflation factor was used to test collinearity (Calcagno 2013). All analyses were performed in R (version 4.0.0).

Carbon and nitrogen properties of vegetation and soil at different successional stages
The carbon and nitrogen contents of soil and vegetation were significantly different between different successional stages (Table 1). Along the three succession stages from SWM to AIM to STM, soil invertase activity increased, while SOC, soil microbial biomass carbon, and vegetation organic carbon decreased. Among them, soil microbial biomass carbon changed significantly with succession, decreasing from 1541.80 mg·kg − 1 SWM to 314.62 mg·kg − 1 STM. Furthermore, vegetation organic carbon content was significantly higher in SWM than in STM (P < 0.05). Regarding soil and vegetation nitrogen levels, the soil TN, soil NH 4 N, and vegetation nitrogen content tended to decrease over the course of the succession. However, the URE activity was the lowest in AIM and increased in the third stage, reaching its maximum value in STM. The contents of soil NO 3 N and soil microbial nitrogen (SMN) followed the same trend over succession, and the lowest values were observed in the second stage. These results indicate that with the succession of vegetation communities, most of the related soil and vegetation properties (such as carbon and nitrogen contents) decreased, while the activities of two enzymes related to carbon and nitrogen increased.

Change of OTU and α-diversity during succession
The number of unique OTU in AIM and STM increased significantly compared with SWM ( Fig. 2A). Furthermore, OTU increased over the succession from SWM to STM (Fig. 2B), andthe highest number of OTU was found in AIM, which was significantly higher than in SWM (P < 0.05). The diversity changed in different successional stages, and the directions of this change were inconsistent. This shows that the Chao1 index and sobs of fungal diversity in AIM and STM increased significantly compared to SWM (Fig. 3A, C). The Shannon index was not significantly different between successional stages (Fig. 3B). These results show that soil fungal OTU increased over succession in the first stage but then reached a relatively stable state in the second successional stage.

Soil microbial community composition in different successional stages
Non-metric multidimensional scaling (NMDS), based on Bray-Curtis distances, showed that the community was separated between different successional 1 3 Vol:. (1234567890) stages (Fig. 4). Further ANOSIM tests (Bray-Curtis) also showed that the distance between each group was significantly larger than that within groups, indicating that significant differences in fungal community structure exist among different succession stages (Fig. S1). Moreover, the result of the ADONIS test showed that the fungal community varied significantly among various successional stages based on Bray-Curtis distance (Table S1).
Fungal communities were characterized by sequencing the V3-V4 hypervariable region of the ITS region. Across all soil samples, 8 phyla, 10 genera, and 10,710 OTU were identified in the soil fungal community. Dominant phyla across all successional  stages were Ascomycota, Basidiomycota, and Mortierellomycota, accounting for more than 74.42% of the fungal sequences from each of the soil samples. In addition, Mucoromycota, Glomeromycota, and Chytridiomycota were present in most soils but at relatively low abundances (Fig. 5A). Two further and rarer phyla Rozellomycota and Neocallimastigomycota were identified in STM only. Ascomycota was the dominant phylum, which was significantly enriched in STM (Fig. 5B). However, Mortierellomycota and Mucoromycota had higher relative abundance in SWM than in other stages (Fig. 5D, E). LEfSe analysis showed that this succession significantly altered fungal communities. Specifically, at the phylum, class, and genus levels, certain fungal groups were significantly enriched at different stages of succession (Fig. 6A). Most fungal taxa were mainly enriched in SWM. For example, Trichocomaceae, Eurotiomycetes, Eurotiales, Penicillium, Aspergillus, Mortierellomycota, Microscypha, Hyaloscyphaceae, Helotiales Hypocreales, Saccharomycetales, and Saccharomycetes increased significantly in SWM (Fig. 6B). Further analysis showed that Ascomycota, Basidiomycota, Mortierellomycota, Mucoromycota, Glomeromycota, Chytridiomycota, and Geminibasidiomycetes followed a linear relationship with habitat change (Fig. S2).
Changes of functional groups of soil fungal communities at different successional stages Based on the abundant information on OTU, the Trophic database annotating fungi and soil fungi was divided into The cladogram shows significant differences between fungal (A) enrichment groups. LDA score chart show biomarker in different succession stages and the length of the histogram represents the influence of different species (B). Taxa with significant differences in abundance between different successions are represented by colored dots, and Cladogram circles represent phylogenetic taxa from phylum to genus. Only the LDA score > 4 for fungi was shown three trophic modes: Pathotroph, Symbiotroph, and Saprotroph (Fig. 7A). The abundance of Pathotroph changed rapidly with succession, and this mode was enriched in SWM and AIM. Symbiotroph did not change significantly, while Saprotroph increased with successional stages. More detailed information on the functional group in soil samples was obtained by FUNGuild. The relative abundance of the fungal functional guilds group changed significantly over the succession. The relative abundances of Animal Pathogen and Endophyte showed parabolic change over the succession, and the highest abundance was found in AIM. However, Ectomycorrhizal and Plant Saprotroph followed the same trend, showing line decreases and reaching the lowest levels of abundance in STM. It is worth noting that the abundance of Plant Pathogen followed a linearly decreasing trend over succession (Fig. 7B). Further, NMDS analysis based on Bray-Curtis distance was employed to assess the structure of functional guild groups. The differentiation of the fungal functional group community was distinct between each stage (Fig. S3).
Carbon factors affecting microbial community structure and function RDA was employed to clarify the relationships between fungal gene changes and carbon and nitrogen levels over succession. Across the first two canonical axes, RDA explained 69.25% of the relationship between fungi at the phylum level with carbon and nitrogen factors (Fig. 8A). Based on model selection, NH 4 N, vegetation nitrogen, and SOC explain the fungal community change best (Fig. 8A, Table S2). The contribution of carbon and nitrogen variables to the variation of the fungal community is illustrated with a modified variation partitioning diagram (Fig. 8B). The complete set of all factors explained 71.61% of the variation in fungal communities of soils, where carbon properties clearly contributed the most (37.28%), exceeding the contribution of nitrogen (30.83%). However, 28.39% of existing fungal community variations cannot be explained by existing factors. The abundance of functional guild groups exhibited an orderly pattern throughout succession. These functional groups were mainly significantly correlated with soil microbial carbon and SOC contents, and had high connectivity with these group nodes. Similarly, in the relationship between fungal functional groups and carbon factors, RDA explained 53.86% (Fig. 9A). The contribution of nitrogen to fungal community variation is higher than that of carbon, and the interaction of carbon and nitrogen explains 14.29% (Fig. 9B). SOC is the main factor affecting fungal community function.

Changes in carbon and nitrogen over alpine grassland succession
The succession of alpine grassland in the Qilian Mountain region, China, provides a dynamic landscape for studying the patterns of microbial   . Other studies showed how soil physicochemical properties change with succession and reported nutrient and soil fertility during the succession (Cong et al. 2015;Jia et al. 2005). It is generally assumed that ecological succession reflects how biological communities reassemble and stabilize following natural or anthropogenic disturbances . However, under the background of climate change, the grassland ecosystem of the Qinghai-Tibet Plateau is facing significant global warming, which may lead to its transformation from an alpine meadow ecosystem to an arid and semi-arid grassland . At present, the carbon content from the stable carbon pool in the alpine meadow decreases, but predicting the extent of this decrease remains challenging (Yuan et al. 2020). Similar to the change of carbon, the nitrogen content also decreases over the succession, except for SMN, NO 3 -N, and URE. Similar results were obtained by Liu et al. (2020) who suggested that soil and environmental factors achieve a stage of stable coexistence when secondary succession develops to a specific stage. However, STM is not the last stage of alpine grassland succession. Existing research showed that the nitrogen content is vulnerable to plant species losses and water stress, but the specific direction of the nitrogen content over succession remains unclear (Barré et al. 2017). Therefore, in the future, more attention should be directed to the succession trend of alpine grasslands.
Taxa-specific changes in soil microbial communities over alpine grassland succession In this study, compared with SWM, OTU sharply increased in AIM and STM, which could be related to water stress and soil nutrients (Shen et al. 2013).
The results are consistent with those of Andreote et al. (2014), who reported that in salt marshes, the ITS gene number increased along succession. The present study also found that OTU of grassland soil fungi tended to stabilize at the later stage of succession, and the number of OTU specific to each successional stage showed a similar trend. These results suggest that the fungal community achieves a temporally steady state in response to environmental factors when succession has developed to a specific stage (Darcy et al. 2018). However, a different perspective was suggested previously, namely that microbial communities may be plastic, rather than static, because of soil environmental changes resulting from succession (Xu et al. 2017). Therefore, these results should be discussed with caution. Fungal diversity among different succession stages was analyzed further. The results showed that the Chao1 and sobs indices in AIM and STM increased significantly compared to SWM. These changes may be due to a shift in the growth strategy of fungi. A widely used framework divides microorganisms into eutrophic and oligotrophic types. While microorganisms of the former type inhabit environments with abundant resources and are characterized by high growth and proliferation rates, the latter inhabit resource-poor environments and must concentrate resources for energy and survival (Paluch et al. 2013). This is also consistent with a study by Zhang et al. who reported that soil bacterial community Shannon and ACE indices increased significantly over succession (Zeng et al. 2017). However, in this study, the Shannon index showed no significant difference between successional stages. No changes in diversity have been reported in the succession of burned forest and paddy fields (Ding et al. 2017;Knelman et al. 2015). The results showed that the abundance of fungal species increased over succession, but either none or only few new species emerged during succession.
NMDS analysis identified a significant shift among different successional stages, which may reflect the response of the fungal community to the changes of succession stages (Fig. 4). A previous study attributed this difference to changes in vegetation community, where vegetation generates the close relationship between plant and fungal communities because plants drive the carbon cycle during early soil development and thus also drive the selection of soil fungal communities (Williams et al. 2013). Moreover, litter has been identified as the main factor leading to a difference in microbial community structure between different habitats . This may explain the significant separation of fungal communities among different succession stages.
Fungal composition is also discussed at the phylum level among different succession stages. At the fungal phylum level, Ascomycota, Basidiomycota, and Mortierellomycota were dominant regardless of succession stage, which is generally consistent with prior research ). In addition, Ascomycota was significantly enriched in STM (Fig. 5B). Mortierellomycota and Mucoromycota showed higher relative abundances in SWM than in other stages. In microbial ecology, it is generally assumed that different fungi with different ecological lifestyles maintain ecosystem health and provide ecosystem functions. Soil fungi can be highly specialized in both their ecosystem function and ecological requirements . The division of fungal strategies can also help to understand the results of the present study. For example, Ascomycota have a long evolutionary history, which endowed them with the capacity to break down most organic substances, and they also help to improve plant performance and plant protection against both biotic and abiotic stresses (Viscarra et al. 2022). Therefore, the change in soil quality caused by changes in vegetation types and litter may promote the observed shift in fungal composition. In microbial ecology, the life strategy of r-k selection forms the basis of the research theory of microbial population reproduction, which reflects an adaptation to living conditions (Fierer et al. 2017). In the present study, the relative abundance of Mortierellomycota and Mucoromycota sharply decreased from SWM to AIM (Fig. S2) and plateaued in the last stage, which is a representative r-to k-strategy transformation. This result is consistent with the findings of Sun et al., who suggested that microbial communities tended to shift from r-to k-strategists both at the phylum and genus levels over succession in the Loess Plateau (Sun et al. 2016). However, the life strategy of Ascomycota retains the k-strategy over succession. In conclusion, combined with the results of this study, fungal communities are more responsive to grassland succession; however, because STM is not the last stage of alpine grassland succession, further long-term detection is needed. LEfSe analysis provides new insights into the responses of soil fungi to altered soil niches. In the present study, most fungal taxa were mainly enriched in SWM, indicating that this habitat provides stable niches for the microbial community. It is generally accepted that certain species dominate the community even in a stable community (Doležal et al. 2013). Thus, most fungal taxa that were enriched in SWM provided a further explanation for the decline in fungal diversity.
Functional changes in fungal communities over alpine grassland succession The FUNGuild results identified a significant difference in function at different stages of succession. Functional groups are also selected by different soil and environmental factors, which present the biochemical cycle of these habitats (Crowther et al. 2019). A previous study suggested that fungal functional groups have habitat-specific adaptations, and therefore, the guilds Pathotroph and Saprotroph are more inclined to inhabit nutrient-rich environments (Nguyen et al. 2016). In the present study, the abundance of Pathotroph was enriched in SWM and AIM, while Saprotroph increased with successional stages, which was consistent with previous research. In addition, the decreased content of soil nutrients leads to a decline of available substrate for fungal trophic digestion during succession. This is especially the case for the Pathotroph guild whose hosts are well represented by plants and soil microorganisms. In addition, Ectomycorrhizal and Plant Saprotroph guilds displayed the same trend throughout succession, showing significant decreases at the first stages and stable abundances from AIM to STM. Moreover, Ectomycorrhizal was considered to indicate a change in the nitrogen cycle (Peter et al. 2001). In the present study, with the decrease of Ectomycorrhizal, most nitrogen factors also decreased, which may be explained by the function of Ectomycorrhizal. However, Animal Pathogen and Endophyte showed U-shaped changes over succession, which contrasted with the abundance of Ectomycorrhizal (Fig. 7). The Ectomycorrhizal guild may restrain the activity of other free fungi by enhancing its ability to acquire organic carbon, thus increasing inter-species competitiveness (Peter et al. 2001).
Carbon and nitrogen shape soil microbial communities and function over succession Soil carbon and nitrogen, both indicators of soil nutrients, are the most important influencing factors (Zhang et al. 2016). A recent study showed that changes in critical soil microbial groups and taxa reflect changes in vital soil biogeochemical processes ). Based on the results of RDA (Fig. 8), most carbon and nitrogen factors can be inferred to correlate with the fungal community. This matches the result of Jiang et al. (2021), who suggested that microbial communities are significantly related to soil nutrients. In addition, 37.28% and 30.83% of the variation of the fungal community could be explained by carbon and nitrogen, respectively, indicating that carbon and nitrogen drive the observed changes in the microbial community. This is consistent with the results of McGee et al. (2019), who reported that various forms of inorganic N may 1 3 Vol.: (0123456789) be important for structuring soil microbial biomass and communities. This leads to an improvement of carbonuse efficiency over forest succession, which links the microbial group classification with the properties of soil organic and nitrogen matter. However, the contents of carbon and nitrogen in this study changed over succession. Interestingly, TN, and SOC contents were significantly related to fungal group during succession, which contrasts with other studies. The present study proved that soil microorganisms are associated with continuous changes in labile, transient nutrient pools, as opposed to basic soil edaphic properties. This may be because the development of soil fungi is more dependent on specific substrate types and their quality (Wala et al. 2006). Grassland succession leads to changes in the vegetation community, which then leads to changes in plant litter quality and associated substrate composition. Under this mode of operation, the grassland matrix in each succession stage will also change. Surprisingly, these results also indicate that changes in soil characteristics caused by the successional alpine grassland affect the fungal community and its functional structure, but these changes are not synchronized.

Conclusion
Soil fungal communities, soil fungal functional groups, plant and soil carbon, and nitrogen characteristics change significantly over the succession from SWM to STM. However, these indexes did not follow the same successional trajectory. Moreover, microbial communities also changed significantly with succession, leading to changes in corresponding functions. More importantly, over the course of succession, fungal community and function were mainly driven by soil ammonium nitrogen and organic carbon rather than by other factors. In conclusion, these results enhance the understanding of the relationship between microbial community, microbial function, and habitat succession in the alpine meadow longterm succession.