Metagenomic insights into the alteration of soil N‐cycling‐related microbiome and functions under long‐term conversion of cropland to Miscanthus

Miscanthus spp. show excellent application prospects due to its bioenergy potential and multiple ecological services. Annual N export with biomass harvest from Miscanthus, even without fertilizer supplement, do not reduce soil N levels. The question arises regarding how Miscanthus can maintain stable soil N levels. Metagenomic strategies were used to reveal soil N‐cycling‐related microbiome and their functional contributions to processes of soil N‐cycling based on the comparison among the bare land, cropland, 10‐year Miscanthus × giganteus, and 15‐year Miscanthus sacchariflorus fields. The results showed that, after long‐term cropland‐to‐Miscanthus conversion (LCMC), 16 of 21 bacterial phyla and all the archaeal phyla exhibited significant changes. Soil microbial denitrification and nitrification functions were significantly weakened, and N fixation (NF) was significantly enhanced. The biosynthesis of amino acids, especially alanine, aspartate, and glutamate metabolism, in soil N‐cycling‐related microbiome was dramatically promoted. The genus Anaeromyxobacter contributed largely to the NF process after LCMC. Variations in the soil available potassium, available N, organic C, and total N contents drove a functional shift of soil microbiome from cropland to Miscanthus pattern. We conclude that Miscanthus can recruit Anaeromyxobacter communities to enhance NF benefiting its biomass sustainability and soil N balance.


| INTRODUCTION
Miscanthus is a C 4 rhizomatous perennial grass native to Eastern Asia (Lewandowski et al., 2000), which has been considered excellent energy crop since mid-1980 (Lemus & Lal, 2005) and recently recognized as a candidate plant for ecological restoration of the degraded land (Mishra et al., 2019). Cultivation of Miscanthus can supply a large quantity of lignocellulosic biomass from which clean ethanol fuels could be derived (Thapa, 2015). It can also provide multiple ecological services, such as soil amelioration (Zhao et al., 2020), water and soil conservation (Miriti et al., 2017), wildlife habitat (Stewart et al., 2009), improvement of landscape effects (Evers et al., 2013), an increase in biodiversity (Stewart et al., 2009), etc. Miscanthus has become popular and attracted the extensive attention from scholars in many research fields worldwide. The replacement of annual row crops with perennial grass bioenergy crops is occurring globally in temperate, tropical, and semi-arid regions, which makes Miscanthus a unique system different from the forest, farmland, and grassland ecosystems due to its vast C sink capacity and simple management requirements (Somerville et al., 2010).
N regulates net plant primary production in terrestrial ecosystems because it is required for plant growth (Lambers et al., 1998). The amount of N stored in soil is related to many factors, such as productivity of vegetation, decomposition of organic matter (OM), rainfall input, dry deposition input, N fixation (NF), N emissions and leaching, etc. (Post et al., 1985). Soil N balance can be maintained artificially through the application of industrial N fertilizers in farmland and grassland ecosystems, which means extra economic input (Egli, 2008). However, most Chinese farmers know little about the details (application and multiple ecological services) of Miscanthus mainly due to the uncertain economic values that could be derived and the late start of related research in this field. Hence, they can not finance the maintenance of Miscanthus for a long-term period, including industrial N fertilizers application, even if the governments provide certain subsidies. Questions arise regarding the soil N balance in Miscanthus when annual harvest happens without any fertilizers supplied, which is a significant concern by most researchers worldwide.
Many studies have reported a considerable amount of N exported from Miscanthus caused by the annual biomass removal, primarily when harvested before the senescence of aboveground parts. However, it has been reported that Miscanthus has relatively low N removal rates (Oliveira et al., 2017) compared to the grain crops (e.g., 3.7% of N removal in maize) due to its nutrients re-translocation at the end of the growing season (Dohleman et al., 2012). Hence, an appropriate amount of industrial N fertilizer is needed to compensate for the annual soil N losses to maintain the high successive biomass yields of Miscanthus; otherwise, the significant decrease in soil N levels could be caused even over a short establishment period (Placek et al., 2018;Wang et al., 2017). However, many papers suggest that soil N contents are not affected much after Miscanthus establishment (Brami et al., 2020;Duan et al., 2019;Dufossé et al., 2014;Gregory et al., 2018;Kang et al., 2020;Krol et al., 2019;Xu et al., 2021) or even increased (Chen et al., 2009;Das et al., 2016;Dufossé et al., 2014;Kahle et al., 1999Kahle et al., , 2002Nebeská et al., 2018) to some degree in the absence of N supplement independent of the cultivation years and soil types. We previously found that the annual harvest of a large amount of biomass (around 22.22 t/ ha/year) over a long-term period (> 15 years) slightly increased total N (TN) contents of 0-100 cm soil in the absence of N fertilization (Zhao et al., 2020(Zhao et al., , 2021. Hence, we can conclude that Miscanthus can provide sustainable biomass resources while naturally maintaining soil N balance even without the compensation of industrial N fertilization, which encourages for the large-scale use of Miscanthus. However, how Miscanthus can naturally maintain soil N balance in the absence of N fertilization still largely remains unclear. Plants and soil microbiome have evolved intimate relationships that include the abundant carbon (C) sources and energy derived from root exudates for soil microbiome, and efficient nutrients, for example, N and phosphorus (P) cycling, as well as some beneficial signaling substances from soil microbial communities promoting the plant growth (Chaparro et al., 2012). Cultivation of Miscanthus can increase (Zhao et al., 2021) or decrease (Duan et al., 2019) the diversity of soil microbial communities, increase fungal/ bacterial ratio (Nebeská et al., 2018;Zhao et al., 2021), and enhance soil N and P-cycling functions of soil microbiome (Duan et al., 2019). It is of concern that Miscanthus cultivation can also improve the activities of soil associative N fixation (ANF, reflected by the increased abundance of N fixation gene nifH) in the bulk or rhizosphere soil when compared with that of the cropland (Li et al., 2015;Soman et al., 2018;Wewalwela et al., 2020;Zhao et al., 2021). Moreover, soil denitrification potential is usually mitigated after the conversion of cropland to Miscanthus. The enhanced soil ANF and the mitigated soil N 2 O emission may contribute to some degree to soil N balance in Miscanthus field, which is still not enough to understand the mechanisms behind natural N balance in unfertilized Miscanthus field. It is necessary to uncover the overall situation of soil N-cycling, the dominant N fixers, and the critical factors in driving the shift of soil microbial communities and functions after unfertilized LCMC.
The objectives of this study were to (1) illustrate the overall situation of soil N-cycling, (2) uncover the dominant N fixers in unfertilized Miscanthus field, and (3) reveal the key factors in driving the shift in soil microbial functions when the cropland was converted to Miscanthus. We hypothesized that (1) soil N-cycling will change toward favoring soil N accumulation and retention, (2) Miscanthus can recruit some specific microbial communities to fix N instead of increasing activities of all soil N fixers, and (3) soil organic C (SOC) may be an essential factor in driving the alteration of soil microbial functions.

| Field site and soil sampling
The experiment is based on a previous test platform and the details could be found in our previous study (Zhao et al., 2020(Zhao et al., , 2021. Before sampling, the apparent plant residues were removed. Four independent sterile augers were used to acquire soil samples on July 2, 2020, from four kinds of cropping field, bare land (maintained for 15 years, converted from >30year corn-wheat rotation cropland), cropland (corn-wheat rotation >30 years), 11-year Miscanthus × giganteus (Migi) field, and 15-year Miscanthus sacchariflorus (Mis) field (converted from corn-wheat rotation cropland). Ten 0-20 cm soil cores (4.0 cm in diameter) from one subplot were randomly collected and thoroughly mixed to form one replicate. There were six replicates for determining soil chemical properties and three replicates were randomly selected from the six replicates for the metagenomic analysis. Soil samples were mixed thoroughly after the removal of apparent stones and plant roots and then sieved through a 2 mm mesh. The thoroughly mixed soil samples were divided into three subsamples. One subsample was transported back to the laboratory in a 4°C incubator and stored in a 4°C refrigerator for NO 3 − -N and NH 4 + -N analysis.
Another subsample was quickly frozen in the steel stainless tank filled with dry ice, transported back to laboratory, and stored in a −80°C refrigerator for metagenomic analysis. And the last subsample was brought back to laboratory under natural conditions, and naturally air-dried to determine other soil properties.

| Soil chemical properties analysis
Around 400 g of soil was carefully prepared before analysis of soil chemical properties. Soil conductivity (SC) and soil moisture (SM) were determined using a WET Sensor Kit (Delta-T Devices Ltd.). Soil pH was measured using a Nahita pH meter, model ST 5000 (OHAUS), after preparing a 1:2.5 ratio of 0.01 M CaCl 2 solution/soil suspension. Soil total N (TN) was determined using a Vario Macro CHNS instrument (Elementar Analys ensysteme GmbH) according to the operating instruction. Soil organic C (SOC) was determined according to the method described by Graham et al. (2019). Simply speaking, SOC contents were calculated by deducting soil inorganic C contents from total C contents (Graham et al., 2019). SOC value was multiplied by the van Bemmelen factor (1.724) to generate soil organic matter (OM) contents. Soil total phosphorus (TP) content was determined using a method by Bertheux (1958), and the Atomic absorption method was used to determine the soil TK content (Khreish & Boltz, 1970). The alkaline KMnO 4 oxidation method, Olsen's extract method, and NH 4 OAc extract method were used to determine soil available N (AN), available P (AP), and available K (AK) contents (Dhillon & Dev, 1979). Soil NO 3 − -N and NH 4 + -N contents were determined colorimetrically using a continuous flow analyzer (Skalar Analytical B.V.) based on the 0.01 mol/L CaCl 2 extract after the alkaline persulfate oxidation (Cabrera & Beare, 1993). Soil cationic exchange capacity (CEC) was determined by leaching the soil with KCl, followed by extraction of exchangeable K + by ammonium acetate (Rhoades, 1982).

| Soil DNA extraction, metagenomic sequencing, and bioinformatics analyses
Total genomic DNA was extracted from soil samples using the E.Z.N.A.® Soil DNA Kit (Omega Bio-TEK) according to the manufacturer's instructions. The concentration and purity of extracted DNA was determined with TBS-380 and NanoDrop2000, respectively. DNA extract quality were checked on 1% agarose gel. DNA extract was fragmented to an average size of about 400 bp using Covaris M220 (Gene Company Limited, China) for paired-end library construction. Paired-end library was constructed using NEXTflex™ Rapid DNA-Seq (Bioo Scientific, Austin, TX, USA). Adapters containing the full complement of sequencing primer hybridization sites were ligated to the blunt end of fragments. Paired-end sequencing was performed on Illumina NovaSeq/Hiseq Xten (Illumina Inc.) at Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China) using NovaSeq Reagent Kits/ HiSeq X Reagent Kits according to the manufacturer's instructions. Sequence data associated with this project have been deposited in the NCBI Short Read Archive database (Accession Number: PRJNA821611).
The raw reads from metagenome sequencing were used to generate clean reads by removing adaptor sequences, trimming and removing low-quality reads (reads with N bases, a minimum length threshold of 50 bp, and a minimum quality threshold of 20) using the fastp (Chen et al., 2018) on the free online platform of Majorbio Cloud Platform (cloud.major bio.com). These high-quality reads were then assembled to contigs using MEGAHIT (Li et al., 2015) (parameters: kmer_min = 47, kmer_max = 97, step = 10), which makes use of succinct de Bruijn graphs. Contigs with the length being or over 300 bp were selected as the final assembling result.
Open reading frames (ORFs) in contigs were identified using MetaGene (Noguchi et al., 2006). The predicted ORFs with the length being or over 100 bp were retrieved and translated into amino acid sequences using the NCBI translation table.
A nonredundant gene catalog was constructed using CD-HIT (Fu et al., 2012) with 90% sequence identity and 90% coverage. Reads after quality control were mapped to the nonredundant gene catalog with 95% identity using SOAPaligner (Li et al., 2008), and gene abundance in each sample was evaluated.
Representative sequences of nonredundant gene catalog were annotated based on the NCBI NR database using blastp as implemented in DIAMOND with e-value cutoff of 1e −5 using Diamond for taxonomic annotations. Cluster of orthologous groups of proteins (COG) annotation for the representative sequences was performed using Diamond (Buchfink et al., 2015) against the eggNOG database (version 4.5.1) with an e-value cutoff of 1e −5 . The KEGG annotation was conducted using Diamond (Buchfink et al., 2015) against the Kyoto Encyclopedia of Genes and Genomes database with an e-value cutoff of 1e −5 .

| Statistical analysis
The distribution and homoscedasticity of data were checked using Shapiro-Wilk and Levene tests before data analysis. The significant differences, at p < 0.05, in soil chemical properties, relative abundances of soil Ncycling-related (NCR) enzymes based on gene mapping, the functional proportion based on the KEGG-pathway-level3 of soil NCR microbial communities, and the relative abundances of soil N fixers among the four groups were evaluated by one-way ANOVA using the least significant difference method corrected by the "Bonferroni" method using the "agricolae" package in R software (3.5.2). The considerable differences in the relative abundances of different phyla between randomly selected two groups were verified using a Wilcoxon rank-sum test corrected by the "Bonferroni" method, and the confidence interval was calculated using the "bootstrap" method. Veen diagrams were drawn using the package "VennDiagram" in R software (3.5.2). The package "vegan" was used to conduct the principal component analysis (PCA), and the "ANOSIM" analysis was carried out to verify the differences among the four groups. The ordinal regression analysis between phylum taxa (based on βdiversity) of soil NCR microbial communities and their functions (KEGG-pathway-level3) was conducted using the package "vegan" in R (3.5.2). The package "vegan" was used to draw the heatmaps of the relative abundances of soil NCR microbial communities and N fixers in R software (3.5.2). One-way ANOVA analysis based on the Tukey-Kramer post-hoc test corrected by the "Bonferroni" method was conducted to verify the multigroup differences in soil NCR phyla at p < 0.05, 0.01, and 0.001, respectively. The LEfSe software was used to generate the bacterial species hierarchy map based on the allagainst-all strategy to show the critical differences-caused microbial communities according to the standard of the linear discriminant analysis (LDA) score greater than 3.5. The relative abundances of the NCR enzymes were mapped based on the KEGG annotation results. The software Circos-0.67-7 was used to generate a Circos diagram to reflect the functional composition (KEGG-pathway-level3) of different groups. The detrended correspondence analysis was initially conducted to evaluate the gradient size of the species distribution, and we found that the data were linearly distributed and the length of the gradient was <3.0, which indicated that the best-fit mathematical model for the data was RDA. The Bray-Curtis method and the Monte Carlo permutation test with 999 random permutations were used to finish the ANOSIM analysis to distinguish the differences in soil microbial composition and functions among the four groups. The package "vegan" was used in R software to draw the RDA figures. The Person's method was used to conduct the correlation analysis using the package "psych" in R software (3.5.2).

| Soil physical and chemical properties of the four kinds of cropping field
As shown in Table 1 (Table 1). When compared with Miscanthus fields, there were not many differences in the mentioned soil indicators except for soil AK, which exhibited significantly higher contents in Mis field than that value in Migi field ( Table 1). The bare land showed a poor soil nutrients status compared to the cropland and the Miscanthus fields (Table 1).

| Changes of soil NCR microbiome after the LCMC
After the sequence spliced, we acquired a total of 7.5 × 10 5 -1.2 × 10 6 contigs and 8.9 × 10 5 -1.4 × 10 6 open reading frame, (ORFs) which averaged 9.6 × 10 5 and 1.2 × 10 6 , respectively (Tables S1 and S2). According to the Venn analysis of soil NCR genes, the bare land had the lowest gene number compared to the other three types of fields, among which the cropland had a slightly larger gene number than the other two types of fields ( Figure 1a). The separated and distanced data points reflected the significant differences in gene patterns among the four groups (Figure 1b). The x, y, and z axes together interpreted around 41.38% of the total variance (Figure 1b). The results of ordinal regression analysis (Figure 1c) showed the significant correlation relationships between the βdiversity of the microbial communities (phylum level) and KEGG-Pathway-Level3 functions (ko00910: N metabolism). In addition to N metabolism, the soil NCR microbial communities also performed the function of alanine, aspartate, and glutamate metabolism, followed by arginine biosynthesis, biosynthesis of amino acids, etc. (Figure 2). The CCTM significantly increased the percentage of the functions of alanine, aspartate, and glutamate metabolism and biosynthesis of amino acids (Figure 2). The bare land had a relatively low percentage of the two dominant functions (Figure 2). The soil NCR microbial communities primarily included bacterial communities and some archaeal communities ( Figure 3). Venn analysis showed that the bare land, cropland, Migi, and Mis field had 23, 12, 20, and 16 unique genera ( Figure S1). The bacterial phyla Actinobacteria and Proteobacteria dominated in the soil NCR microbial communities, which accounted 56.10%-70.80% of the total abundances, followed by the phyla Chloroflexi, Acidobacteria, Firmicutes, Thaumarchaeota, Planctomycetes, Nitrospirae, and Verrucomicrobia ( Figure 3). The CCTM increased the relative abundances of the dominant phyla Actinobacteria and Proteobacteria, especially when compared with the bare land ( Figure 3). The relative abundances of the other phyla mainly were not affected much or decreased to different degrees while decreasing the relative abundances of the other phyla ( Figure 3). Except for the top three dominant phyla, the relative abundances of all the other phyla in bare land exhibited significantly higher than those of the other three types of fields (Figure 3).  Linear discriminant analysis (LDA) effect size (LefSe) analysis (LDA >3.5, p < 0.05) revealed the enriched bacteria (from phylum to genus levels) with significantly different abundances in the four kinds of cropping fields (Figure 4a,b). At the bacterial phylum level, the phyla Actinobacteria and Proteobacteria enriched in Miscanthus F I G U R E 2 Collinearity analysis of soil microbial functions in the four kinds of cropping field. Migi and Mis represented 15-year Miscanthus sacchariflorus and 10-year Miscanthus × giganteus, respectively. The "Cropland" represented the normal farmland for wheat and maize rotation and the "Bareland" represented the land without any plant growth as described in materials and methods part. The data following each function represented the averaged proportion of function, and the four groups were arranged in the order of Bareland, Cropland, Migi, and Mis field from the left to the right. The small different letters following the data indicated the significant differences at p < 0.05 in the proportion of different functions among the four groups. The "Cropland" represented the normal farmland for wheat and maize rotation and the "Bareland" represented the land without any plant growth as described in materials and methods part. The Arabic numbers (1, 2, and 3) behind Migi, Mis, Bareland, and Cropland represented the three single replicate samples for each group. The significant differences at p < 0.001 among the four groups were tested based on the ANOSIM analysis. fields, respectively, which whereas were significantly depleted in the bare land and cropland (Figure 4a,b). At the bacterial family level, the families Solirubrobateraceae and Geodermatophilaceae were the enriched bacteria in Migi field, and the families Microbacteriaceae and Bradyrhizobiaceae were the enriched bacteria in Mis field (Figure 4a,b). The Migi field was more prevalent at the genus level than the other three kinds of cropping fields with the genus Solirubrobater. And the Mis land was more dominant at the genus level than the other three kinds of fields with the genus Bradyrhizobium and Agromyces (Figure 4a,b). For the Archaea, the phylum Euryarcheota dominated in the Mis field after LCMC. At the genus level, Methanomassiliicoccus and Methanocella enriched in the Miscanthus fields, respectively, after LCMC ( Figure S2).

| Changes in soil N-cycling and the dominant contributors
The overall network diagram of the N cycle of the four kinds of cropping fields manifested that N cycle status varied conspicuously after LCMC ( Figure 5). Summarily speaking, the N fixation process was enhanced dramatically ( Figure 5). The L-glutamine to L-glutamate step was promoted significantly ( Figure 5). Nevertheless, the denitrification processes were entirely diminished ( Figure 5). Moreover, the critical step from ammonia to hydroxylamine of conversion of ammonia to nitrate process was likewise debilitated ( Figure 5). In addition, the conversion of soil-toxic materials nitroalkane and nitrile to nitrite and ammonia, respectively, were both undermined ( Figure 5). The results of functional contribution analysis evidenced almost the same microbial participants in different soil N-cycling processes whereas the different proportions of the relative contribution at the genus level ( Figure 6). The genus Anaeromyxobacter was the principal contributor to the N fixation function, followed by the genus Hyphomicrobium (Figure 6), which was also evidenced by the increased proportion (based on the total functions) of N fixation function and the relative abundances of these two genera ( Figure S3; Table S3). For the nitrification, the genus unclassified_p_ Crenarchae followed by Nitrososphaera, Streptomyces, and Nocardioides, accounted more than 90% of function 23 (1.14.99.39). Nevertheless, the genus unclassified_p__ Candidatus_Rokubacteria only performed function 28 (1.7.2.6) in the cropland and bare land soil but almost did F I G U R E 3 Heatmap of the relative abundances of soil N-cycling-related bacteria and archaea at the phylum level. Migi and Mis represented 15-year Miscanthus sacchariflorus and 10-year Miscanthus × giganteus, respectively. The "Cropland" represented the normal farmland for wheat and maize rotation and the "Bareland" represented the land without any plant growth as described in materials and methods part. The coloring was based on the logarithmic proportion of each phylum and the transition from the red color to the blue color indicated the gradual decrease in the relative abundances of phyla. The Arabic numbers (1, 2, and 3) behind Migi, Mis, Bareland, and Cropland represented the three single replicate samples for each group and the significant differences at p < 0.05 in the relative abundances of the microbiome among the four groups were marked by the different small letters.

F I G U R E 4
Differential discriminant analysis of soil bacteria revealed by (a) the species hierarchy map from the phylum to genus level and (b) the LDA discriminant columns (>3.5). Migi and Mis represented 15-year Miscanthus sacchariflorus and 10-year Miscanthus × giganteus, respectively. The "Cropland" represented the normal farmland for wheat and maize rotation and the "Bareland" represented the land without any plant growth as described in materials and methods part. The circle with yellow color implied no significant differences between two certain groups, and the other colors represented the significant differences between any two groups among the four groups.

F I G U R E 5
Functional map displayed in the form of the related enzymes based on the relative abundances of soil N-cycling-related genes. The significant differences at p < 0.05 among the four groups were marked by the different small letters sequenced in the order of bare land, cropland, Migi, and Mis field from the left to the right. Migi and Mis represented 15-year Miscanthus sacchariflorus and 10-year Miscanthus × giganteus, respectively. The "Cropland" represented the normal farmland for wheat and maize rotation and the "Bareland" represented the land without any plant growth as described in materials and methods part. The color from the bright green to the red represented the gradual increase in the relative abundances. not work in the two Miscanthus fields (Figure 6). The relative contributions of different soil microbial communities to the soil denitrification process exhibited few differences after the changes in land-use types ( Figure 6). There were some changes in the relative contribution of soil microbiome to the dissimilatory nitrate reduction, such as the decrease in the genera Nitrospira and Nitrososphaera, and also to dissimilatory nitrate reduction process, such as the increase in genera Bradyrhizobium and Solirubrobacter (Figure 6), and the changes were also well evidenced in the proportion based on the total functions (Table S4).

| Interaction analysis between soil microbial functions and soil properties
The correlation heatmap displayed that soil pH, SOC, TN, AN, AK, NN, and SM were the primary factors that significantly affected the functions and relative abundances of soil microbiome (Figure 7; Figure S4). SOC, TN, AN, and AK exhibited extremely positive correlations with most soil microbial functions, including N fixation (Figure 7), and with the relative abundances of the dominant phyla Actinobacteria, Proteobacteria, and Acidobacteria ( Figure S4). However, soil pH, NN, and SM negatively affected most the soil microbial functions (Figure 7) and the relative abundances of the dominant phyla Actinobacteria, Proteobacteria, and Acidobacteria ( Figure S4). RDA analysis presented soil microbial groups based on the functional analysis of the bare land, cropland, and Miscanthus fields was distributed in different quadrants. Soil AK, SOC, AN, and TN were significantly correlated with the soil microbial functions (Figure 7) and structures ( Figure S4) in Miscanthus fields. Soil NN, SM, pH, AP, and TP were significantly correlated with soil microbial functions in cropland (Figure 7).

| Effects of the LCMC on soil properties
The expected high biomass yield and difficultly accepted fertilizer uses make Miscanthus a unique system that F I G U R E 6 Relative contributions of different soil microbial communities at genus level to different processes of soil N metabolism. Migi and Mis represented 15-year Miscanthus sacchariflorus and 10-year Miscanthus × giganteus, respectively. The "Cropland" represented the normal farmland for wheat and maize rotation and the "Bareland" represented the land without any plant growth as described in materials and methods part. differs distinctively from the farmland, grassland, wetland ecosystems, etc. A great anxiety due to a knowledge gap in the impacts of the annual harvest of unfertilized Miscanthus on soil quality, which severely impedes its large-scale uses, shows the necessity of this research. In this study, we examined the alterations of some selected soil properties, composition, and functions of soil NCR microbiome. We importantly revealed the whole N-cycling and the N fixers in Miscanthus. The long-term cultivation of Miscanthus can induce a risk of soil acidification due to the exudates from the developed Miscanthus root system (Hu et al., 2018), which was also detected in this study according to the significantly decreased soil pH (Table 1). The relatively low SM in Miscanthus fields compared with the cropland and bare land (Table 1) involved with the blocking effects of the dense stems and leaves on a light rain shortly before the sampling.
A plethora of papers has reported the substantial C sequestration effects after Miscanthus cultivation independent of climate and soil types, Miscanthus species, and management practices, as well as the former landuse history (Clifton-Brown et al., 2007;Zang et al., 2018). In this work, the significantly increased SOC and OM in Miscanthus fields compared with the cropland and bare land (Table 1) forcefully confirmed the outstanding C accrual effects even when the biomass was harvested annually without any fertilizers applied. We believe that the stubble residues after annual harvest, root exudates, and turnover were essential reasons. In addition, the slower SOC mineralization due to the absence of soil tillage should also be considered, (Soussana et al., 2004) although the actual effect of soil tillage on SOC stocks is questioned (Powlson et al., 2014).
Dozens of published papers have reported on the elements exported from Miscanthus field by biomass harvest, suggesting the appropriate supplement of the necessary elements (Dufossé et al., 2014;Oliveira et al., 2017;Singh et al., 2015;Yost et al., 2018). We herein observed the significantly decreased TP and AP (Table 1), which further evidenced the need for P additions after the repeated F I G U R E 7 (a) Heatmap of the correlation between different functions and different soil factors, and (b) the redundancy analysis of the soil microbial functions restricted by soil factors. SOC, TN, AN, AK, NN, CEC, TP, AP, AnN, and SM represented soil organic C, total N, available N, available potassium, nitrite N, cation exchange capacity, total phosphorus, available phosphorus, ammonium N, and soil moisture, respectively. Migi and Mis represented 15-year Miscanthus sacchariflorus and 10-year Miscanthus × giganteus, respectively. The "Cropland" represented the normal farmland for wheat and maize rotation and the "Bareland" represented the land without any plant growth as described in materials and methods part. *, **, and *** indicated the significant correlation relationships at p < 0.05, 0.01, and 0.001, respectively. The different colors indicated the correlation coefficients as noted by the color bar.
harvests (Oliveira et al., 2017). However, the slight decrease in soil TK and significant increase in soil AK in Miscanthus fields (Table 1) reflect the dispensable complement of K fertilizer after the successive biomass harvest in Miscanthus, which was also consistent with our previous conclusions (Zhao et al., 2020(Zhao et al., , 2021. We speculated that the accrual effects of stubbles left after the annual harvest alleviated the depletion of TK in the top soil layer (0-30 cm), and the vast K stocks in soil could sustain the yearly biomass production even without any fertilizer. The significant decrease in SC after LCMC (Table 1) was primarily attributed to the significantly decreased soil inorganic ions such as NO 3 − , NH 4 + , etc., (Table 1) caused by the annual harvest, which was also detected in a 2-year Miscanthus field (Kang et al., 2013). However, we did not find significant differences in soil CEC among the four groups (Table 1), which was probably due to the slight influence of Miscanthus cultivation on soil micro-elements such as K + , Na + , Ca 2+ , and Mg 2+ . The differences in Miscanthus species and establishment years, management practices, application of N fertilizer, harvest time and frequency, and soil status complicated the effects of Miscanthus cultivation on soil N levels. The conclusions were still controversial (Arundale et al., 2014;Dufossé et al., 2014;Zhao et al., 2020). We previously found that the annual N export by harvest slightly increased soil N levels instead of N depletion compared with the preplanting levels even without any fertilizers over a long-term (> 12 years) period (Zhao et al., 2020(Zhao et al., , 2021. One of the most critical findings in this study is the significant increase in soil TN and AN after LCMC (Table 1), which is slightly higher than the preplanting value (Zhao et al., 2020). Hence, we first and strongly evidenced that Miscanthus could naturally maintain soil N balance even when the N element was annually removed by the harvest in the absence of any fertilizers, which also reflected the shift in soil microbial functionalities. Moreover, the lack of N fertilizer and the root absorption induced the significantly decreased soil NN in Miscanthus field, which indicated a relatively low N leakage risk compared with that of the cropland and bare land.

| Effects of the LCMC on soil microbiome
The activation and enrichment effects of Miscanthus cultivation on the soil microbiome, especially some beneficiary microbiome, has been widely confirmed independent of Miscanthus species, stand age, and management practices, as well as the soil and climate types Li et al., 2016;Wu et al., 2022;Zadel et al., 2022). However, alterations of soil NCR microbial composition caused by long-term unfertilized Miscanthus cultivation are poorly understood. In this work, long-term unfertilized Miscanthus drove the shift of soil microbiome from cropland to Miscanthus pattern (Figures 1a,b and 3), where the increased relative abundances of the dominant phyla Actinobacteria and Proteobacteria, coupled with the decreases of the low-relative-abundance phyla, for example, Crenarchaeota, Gemmatimonadetes, and Nitrospirae were primary factors in structure reshaping of soil NCR microbiome (Figures 3 and 4a,b). The slight differentiation in soil NCR microbial patterns between Mis and Migi fields indicated the different effects of different Miscanthus species and stand ages on soil NCR microbiome (Figures 1a,b  and 3). Substantially different from cropland, Mis, and Migi field, bare land had remarkably different soil NCR microbial structure patterns, characterizing as the significantly lower dominant phyla abundances (Figure 3), implying the prominent enrichment effects of Miscanthus on dominant phyla of soil NCR microbiome.
Miscanthus cultivation can enhance microbial functions of soil N-cycling, facilitating N absorption and improving the N using efficiency of Miscanthus (Ma et al., 2021). Interestingly, several reports have evidenced that Miscanthus cultivation, even in the first year (Keymer & Kent, 2014), can recruit the N fixers to provide nonnegligible N to meet Miscanthus growth demand (Kane et al., 2022;Zhao et al., 2021), especially under N deficiency conditions (Liu & Ludewig, 2019). We herein, together with our previous report (Zhao et al., 2021), strongly supported this point in different Miscanthus species and stand ages (10 and 15 years) based on a LCMC. Whereas few documents mentioned which communities were the primary N fixers in the Miscanthus system, which was still a knowledge gap for research communities working on Miscanthus. Our results manifested that the genus Anaeromyxobacter was the principal N fixer in the Miscanthus system in Northern China ( Figure 6; Figure S3 and Table S3). Additionally, the genus Sinorhizobium seemed to play a noteworthy role in the N fixation process in Migi field ( Figure S3). These results provided a theoretical basis for further study on the mechanisms of N fixation in Miscanthus and a good example for the research of N fixation in graminaceous crop, which is of great significance for low-fertilization-input agricultural production.
Excessive N fertilization usually induces enhanced soil nitrification in cropland, which renders N leakage into groundwater or surface runoff into the adjacent water, threatening people's health (Guo et al., 2014). In this work, the substantially weakened soil microbial functions related to nitrification ( Figure 5) signified the reduced potential of soil N leakage and surface runoff in Miscanthus field regardless of the conversion period. Additionally, the weakened soil denitrification processes ( Figure 5) indicated the reduced potential of soil N 2 O emission, which is a greenhouse gas with 300-fold more harmful than CO 2 , in Miscanthus fields. The decreased soil nitrification and denitrification processes in Miscanthus fields, primarily caused by the long-term absence of soil N fertilization, were also the nonnegligible factors for N retention and balance in the soil of Miscanthus fields. Hence, LCMC turned the environment-unfriendly N cycle driven by the excessive industrial N into an environment-friendly pattern driven by the soil biological N fixation, inspiring for the government or companies during the large-scale utilization of Miscanthus in the ecological restoration of degraded land. Soil N cycle and the microbiome are generally closely related to some primary soil properties, for example, soil NCR genes were evidenced to be negatively correlated with soil properties related to acidity in agriculture to forest conversion in the Amazon region (Merloti et al., 2019). In this work, among the measured soil properties, soil AK, SOC, AN, and TN were revealed to play pivotal roles in driving no matter the functions or the structure patterns of soil NCR microbiome (Figure 7; Figure S4) after LCMC. While soil NiN played some critical roles in shaping the functional and structural patterns of soil NCR microbiome in the cropland.

| Conclusions
In this work, the LCMC has considerably altered soil properties, affecting the composition and function of soil NCR microbiome in Northern China. In this manner, SOC and NiN were the main soil chemical properties in differentiating the structural and functional patterns in cropland and Miscanthus fields. Our results also demonstrated that the soil N fixation process was significantly enhanced where the genus Anaeromyxobacter served as the dominant N fixers, and soil nitrification and denitrification processes were weakened dramatically after LCMC. The enhanced N retention due to the decreased N leakage and emission potential and increased N fixation potential in the soil of Miscanthus fields contributed to the N balance during such a long period (15 years). Taken together, our results clarify the effects of LCMC on the compositional and functional characteristics and may aid in future research that seeks the mechanisms of how Miscanthus recruits N fixers to realize the sustainability of soil N and biomass production.

ACKNOWLEDGMENTS
This work was supported by Special Projects for Capacity of Scientific and Technological Innovation (KJCX20210419 and KJCX20230107) and Hebei Province Key Research and Development Programs (22326415D and 22324002D).