Trichoderma Biofertilizer Links to Altered Soil Chemistry, Altered Microbial Communities, and Improved Grassland Biomass

In grasslands, forage and livestock production results in soil nutrient deficits as grasslands typically receive no nutrient inputs, leading to a loss of grassland biomass. The application of mature compost has been shown to effectively increase grassland nutrient availability. However, research on fertilization regime influence and potential microbial ecological regulation mechanisms are rarely conducted in grassland soil. We conducted a two-year experiment in meadow steppe grasslands, focusing on above- and belowground consequences of organic or Trichoderma biofertilizer applications and potential soil microbial ecological mechanisms underlying soil chemistry and microbial community responses. Grassland biomass significantly (p = 0.019) increased following amendment with 9,000 kg ha−1 of Trichoderma biofertilizer (composted cattle manure + inoculum) compared with other assessed organic or biofertilizer rates, except for BOF3000 (fertilized with 3,000 kg ha−1 biofertilizer). This rate of Trichoderma biofertilizer treatment increased soil antifungal compounds that may suppress pathogenic fungi, potentially partially responsible for improved grassland biomass. Nonmetric multidimensional scaling (NMDS) revealed soil chemistry and fungal communities were all separated by different fertilization regime. Trichoderma biofertilizer (9,000 kg ha−1) increased relative abundances of Archaeorhizomyces and Trichoderma while decreasing Ophiosphaerella. Trichoderma can improve grassland biomass, while Ophiosphaerella has the opposite effect as it may secrete metabolites causing grass necrosis. Correlations between soil properties and microbial genera showed plant-available phosphorus may influence grassland biomass by increasing Archaeorhizomyces and Trichoderma while reducing Ophiosphaerella. According to our structural equation modeling (SEM), Trichoderma abundance was the primary contributor to aboveground grassland biomass. Our results suggest Trichoderma biofertilizer could be an important tool for management of soils and ultimately grassland plant biomass.

In grasslands, forage and livestock production results in soil nutrient deficits as grasslands typically receive no nutrient inputs, leading to a loss of grassland biomass. The application of mature compost has been shown to effectively increase grassland nutrient availability. However, research on fertilization regime influence and potential microbial ecological regulation mechanisms are rarely conducted in grassland soil. We conducted a two-year experiment in meadow steppe grasslands, focusing on above-and belowground consequences of organic or Trichoderma biofertilizer applications and potential soil microbial ecological mechanisms underlying soil chemistry and microbial community responses. Grassland biomass significantly (p = 0.019) increased following amendment with 9,000 kg ha −1 of Trichoderma biofertilizer (composted cattle manure + inoculum) compared with other assessed organic or biofertilizer rates, except for BOF3000 (fertilized with 3,000 kg ha −1 biofertilizer). This rate of Trichoderma biofertilizer treatment increased soil antifungal compounds that may suppress pathogenic fungi, potentially partially responsible for improved grassland biomass. Nonmetric multidimensional scaling (NMDS) revealed soil chemistry and fungal communities were all separated by different fertilization regime. Trichoderma biofertilizer (9,000 kg ha −1 ) increased relative abundances of Archaeorhizomyces and Trichoderma while decreasing Ophiosphaerella. Trichoderma can improve grassland biomass, while Ophiosphaerella has the opposite effect as it may secrete metabolites causing grass necrosis. Correlations between soil properties and microbial genera showed plant-available phosphorus may influence grassland biomass by increasing Archaeorhizomyces and Trichoderma while reducing Ophiosphaerella. According to our structural equation modeling (SEM), Trichoderma abundance was the primary contributor to aboveground grassland biomass. Our results suggest Trichoderma biofertilizer could be an important tool for management of soils and ultimately grassland plant biomass.
Keywords: in situ fertilization experiment, high-throughput sequencing, soil chemistry, key fungal genera, structure equation modeling INTRODUCTION Grasslands are valuable ecological resources due to their biodiversity and ecosystem functions (Soussana et al., 2007;Hermann et al., 2016). In recent years, grasslands have been over-exploited for grazing with consequences such as soil carbon loss and desertification (Kosmas et al., 2014;Conant et al., 2017). Aboveground plant biomass is the foundation of grassland animal husbandry, which is often co-limited by both N and P in semi-arid grasslands (Zhan et al., 2017). Globally, many grasslands receive little nutrient input, and grassland fertility declines are linked to reduced grassland provisioning services, such as primary productivity (Sattari et al., 2016). Herbivore manure is important in grassland nutrient cycling; however, intensified livestock production can lead to excess manure. Deposition of this waste can contaminate groundwater and increase nitrate and phosphate concentrations, causing soil hardening and/or salinization, negatively impacting soil physicochemical properties and soil microbial communities (Groenigen et al., 2005;Sun et al., 2015;Xu et al., 2017). To mitigate these issues, manure can be processed into organic fertilizer via composting (Moral et al., 2009;Tang et al., 2011). When applied in grasslands, compost effectively increases soil fertility, supporting soil microbial communities and improving grassland productivity (van der Heijden et al., 2008;Wagg et al., 2014).
Trichoderma spp. are well-known plant growth-promoting fungi (Masunaka et al., 2011) that enhance plant nutrient uptake, production of growth hormones, and protect plants from pathogen infection (De Souza et al., 2008;Contreras-Cornejo et at., 2009;Zhang et al., 2013a). Many plant growthpromoting fungi are effective in vitro, but do not significantly benefit plants in field studies because of their inability to colonize root systems in situ. The survival rate and population density of plant growth-promoting fungi are prerequisites for their effectiveness in the rhizosphere (Raaijmakers et al., 2009;Zhang et al., 2013b). Formulation of inoculum carriers is necessary to ensure survival and maintenance of a viable Trichoderma population in field soils. Combination of organic fertilizers (compost) and Trichoderma strains as biofertilizers (BOF) may be an effective way to facilitate greater plant biomass compared to amendments of organic fertilizers or Trichoderma separately (Huang et al., 2011;Zhang et al., 2013a). Organic matter in BOF improves soil fertility while being an excellent substrate for the growth of Trichoderma. Zhang et al. (2013b) showed application of Trichoderma biofertilizer significantly altered fungal diversity and community in the rhizosphere of cucumber, resulting in improved yields. Additional researches demonstrated Bacillus and Trichoderma biofertilizers improved banana growth and increased soil microbial diversity Fu et al., 2017;Xiong et al., 2017). However, to the best of our knowledge, microbial ecological mechanisms influencing soil microbial genera and subsequent stimulation of root exudation, which supports plant growth responses to biofertilizers are rarely researched in grasslands.
Rhizosphere chemistry is the sum of root exudates, their breakdown products, and the microbial products of soil-derived chemicals (Pierre et al., 2017). Based on this, we defined extracted chemicals in bulk soil as "soil chemistry." Grasslands harbor diverse plant species with diverse root exudates. Soil chemistry composition plays an important role mediating plant and soil microbial interactions (Oburger and Schmidt, 2016). Dessaux et al. (2016) demonstrated plant roots convert their associated soil into complex mesotrophic environments, supporting highly diverse microbial communities. Root exudates provide a major food source for soil microbes and directly affect their assemblage and reproduction (Raaijmakers et al., 2009). The studies of Ng et al. (2014) and Salles et al. (2009) also reported soil carbon composition strongly influences soil microbial community composition. Hence, plant roots may drive multitrophic interactions in soil via root exudation. Furthermore, some chemicals with antifungal activities may produce a chemical barrier to pathogenic fungi (Yuan et al., 2012;Raza et al., 2015), thereby indirectly improving grassland biomass. Beneficial plantmicrobe interactions have received less attention in grassland systems, but advances in microbial genetics are improving our ability to link key microbial genera with soil health and ecosystem productivity (Finkel et al., 2017).
In our study, we applied organic fertilizer (composted cattle manure) or Trichoderma biofertilizer (composted cattle manure + inoculum) at different rates (0 kg ha −1 , 3,000 kg ha −1 , 6,000 kg ha −1 , or 9,000 kg ha −1 ) to meadow steppe grassland plots for two growing seasons. We focused on the potential microbial ecological mechanisms underlying observed grassland biomass. This investigation addresses the following questions: (i) How do different fertilization regimes impact grassland biomass? (ii) Does addition of Trichoderma further increase grassland biomass above the effects of cattle manure alone? (iii) Do different fertilization regimes shift soil chemistry and soil microbial communities? (iv) Is soil chemistry significantly correlated with specific microbial genera? (v) Are soil microbial diversities or communities significantly correlated with grassland plant biomass? (vi) How do different variables (i.e., key microbial genera, soil properties, soil chemistry, and soil microbial communities) contribute to grassland plant biomass? Our findings will provide a better understanding of above-and belowground responses to alternative grassland fertilization and may ultimately improve management of grassland soils for key provisioning services.

Trichoderma Strain, Culture Conditions, and Conidia Suspension
We utilized a strain of Trichoderma rossicum, NAU-18 (CCTCC No. AF2017008, China Center for Type Culture Collection), isolated from soil at the Grassland Agro-ecosystems Station (49 • 26 ′ 12 ′′ N, 120 • 8 ′′ 52 ′′ E, 695 m altitude). NAU-18 was identified based on morphological characteristics and internal transcribed spacer (ITS) sequence analysis ( Figures S1A,B). The strain was stored at −80 • C in 30% glycerol before use and routinely cultured on potato dextrose agar (PDA) at 28 • C. NAU-18 conidia suspension was prepared according to the procedure of Zhang et al. (2013a). The final concentration was 3.2 × 10 8 colony-forming units (CFU) ml −1 , based on hemocytometer counts. This conidia suspension was used as inoculum for solid-state fermentation.
Wheat straw was utilized as a substrate for NAU-18 production. Sterile wheat straw and 1% urea were stirred thoroughly (moisture = 70%), after which NAU-18 conidia suspension was added to the substrate (9% of the total [v/w]), and solid-state fermentation was maintained at 30 • C for 8 days in a shallow tray. The final dry weight density of NAU-18 was 1.1 × 10 10 CFU g −1 . Biofertilizer (BOF) was a mixture of 1:10 (w/w) Trichoderma NAU-18 fermentation products and composted cattle manure.

Study Site Description and Experimental Design
Our in situ fertilization experiment was initiated in June 2015 at the Grassland Agro-ecosystems Station, located in northeastern Inner Mongolia, Hulunbuir, China. This region is characterized by a continental temperate semi-arid climate typically with 350-400 mm annual precipitation (Inner Mongolia Meteorological service, http://www.imwb.gov.cn). The growing season is from late June to early September with an average temperature range of 16-21 • C. The soil in this region is classified as a chernozem (IUSS WG WRB, 2015) with pH 6.59, 44.5 g kg −1 soil organic matter, 189.3 mg kg −1 plant-available N, 3.3 mg kg −1 plant-available P, and 134.2 mg kg −1 plant-available K. The grassland plant community is dominated by Poapratensis L., Leymus chinensis (Trin.) Tzvelev, Bromusinermis Leyss, Stipabaicalensis Rochev., and Potentilla bifurca L.

Soil and Vegetation Sampling
Soil and vegetation samples were collected on August 25, 2016. Five cylindrical soil cores (0-10 cm depth, 6 cm diameter) were randomly collected and mixed to form a composite soil sample for each plot. All soil samples were sieved through 2.0-mm mesh and thoroughly homogenized. Each composite soil sample was then divided into three sub-samples: the first sub-sample was stored at 4 • C for soil chemistry analysis, another was air dried at room temperature for 7 days for analysis of soil properties, and the final sub-sample was stored at −20 • C for DNA extraction. Two random subplots (0.25 × 1.0 m) within each plot were destructively harvested for aboveground biomass. Shoots were cut at the soil surface and oven dried at 65 • C for 72 h before weighing. Shoot dry weights were expressed as total grassland biomass per m 2 . Plant tissue crude protein was determined by Kjeldahl method (Da Silva et al., 2016).

Extraction and Identification of Soil Chemistry
For each replicate soil sub-sample, 5.0 g of soil were combined with 50 ml of ethyl acetate (1:10 ratio) in a 150 ml Erlenmeyer flask and shaken on a table agitator at 30 • C for 2 h. The suspension was filtered through a 0.45-µm Millipore filter and concentrated to 0.5 ml at 35 • C using a vacuum rotary evaporator (Yarong Model RE-52A, Shanghai, China). The concentrated solution analyzed by gas chromatography-mass spectrometry (GC-MS). The following protocol was used for GC-MS detection: 60 • C for 3 min, followed by increasing the temperature at a rate of 5 • C min −1 to 240 • C, held at 240 • C for 5 min, increased to 280 • C at a rate of 20 • C min −1 , and held at 280 • C for 2 additional minutes. The mass spectrometry detector (MSD) was operated in electron ionization mode at 70 eV, with a source temperature of 230 • C. A continuous scan from 45 to 500 m/z was applied. Helium was the carrier gas at a linear velocity of 1.0 ml min −1 . The chromatographic peaks of the components were compared with entries in the National Institute of Standards and Technology (NIST) database (Version 2.0) to characterize the variation in the chemical composition of soil sub-samples.

Determination of Soil Physicochemical Properties
Soil analyses were performed by the soil-testing lab in Qiyang at the red soil experimental station of the Chinese Academy of Agricultural Sciences. Soil pH was measured with soilwater (1:2.5, w/v) slurry using a compound electrode (PE-10; Sartorious, Germany). Soil organic C (SOC) and total N (TN) were determined with an Elementar Vario EL III (Germany). Plant-available N was hydrolyzed with 1.0 mol l −1 NaOH and measured according to the methods of Shi (1996). Total P was determined by perchloric acid digestion (Olsen and Sommers, 1982). Plant-available phosphorus (Olsen-P) was extracted with sodium bicarbonate and measured using the molybdenum-blue method (Watanabe and Olsen, 1965). Total and plant-available K were measured by flame photometry (Knudsen et al., 1982).

DNA Extraction, High-Throughput Sequencing, and Bioinformatics
Total soil genomic DNA was extracted using a Power Soil DNA Isolation Kit (MoBio Laboratories Inc., Carlsbad, CA, USA) following the manufacturer's instructions. The concentration and quality (A260/A280 ratio) of the DNA samples were determined using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). Pyrosequencing analyses of the 16S rRNA gene and ITS region were performed to determine the diversity and composition of bacterial and fungal communities, respectively. The V4 region of the bacterial 16S rRNA gene was amplified using the genespecific primers 520F (5 ′ -AYTGGGYDTAAAGNG-3 ′ ) and 802R (5 ′ -TACNVGGGTATCTAATCC-3 ′ ), while the ITS1 region of the fungal ITS was targeted by the primers ITS1F (5 ′ -CTTGGTCATTTAGAGGAAGTAA-3 ′ ) and ITS2 (5-GCTGCGTTCTTCATCGATGC-3). PCR amplification of the bacterial 16S rRNA and fungal ITS sequences was conducted in a volume of 30 µl containing15 µl of 2 × Master Mix (Thermo Scientific R Phusion High-Fidelity PCR Master Mix, New England Biolabs, UK), 0.5 µM each primer, 10 ng of soil DNA template and nuclease-free water to bring the total to 30 µl. The obtained PCR products were purified using a PCR Purification Kit (Axygen Bio, USA) and quantified with PicoGreen R dsDNA reagent (Promega, USA). The purified amplicons were then pooled in equimolar concentrations as a single aliquot and employed for library construction using the NEB Next R Ultra TM DNA Library Prep Kit for Illumina (New England Biolabs, UK). All library preparation was performed on the Illumina MiSeq platform at Total Genomics Solutions Biotechnology Co., Ltd. (Shenzhen, China).
Raw sequences were assembled for each sample based on unique barcodes using Quantitative Insights Into Microbial Ecology (QIIME) after removal of the adaptor and primer sequences (Caporaso et al., 2011). Split sequences for each sample were merged using FLASH V1.2.7 (Magoc and Salzberg, 2011), and low-quality sequences were then discarded using QIIME. The sequences retained for each sample were analyzed following the UPARSE pipeline, using USEARCH and Perl scripts to generate an operational taxonomic units (OTUs) table and pick representative sequences (Edgar, 2013). Sequences with a quality score < 0.5 or a length < 200 nt and singletons were discarded. After that, retained sequences were assigned to OTUs at 97% similarity. We chose a representative sequence from each OTU, and the Ribosomal Database Project (RDP) classifier (the RDP Bacterial 16S database for 16S rRNA data and the UNITE Fungal ITS database for ITS data) was used to assign taxonomic information (Kõljalg et al., 2013). The MOTHUR (version 1.25.1) standard operating procedure (SOP) was employed for further analyses of pyrosequencing data (Wang et al., 2007;Schloss et al., 2009). To correct for sampling effects, randomly selected subsets of 27,942 sequences per sample for 16S and 30,629 sequences per sample for ITS were chosen for further bacterial and fungal community analyses.

Statistical Analyses and Data Accessibility
Statistical analyses of all parameters were performed using the IBM SPSS statistical software package version 20 (IBM Corporation, New York, USA). Data from each treatment were analyzed using one-way analysis of variance (ANOVA), and Duncan's multiple range tests (p < 0.05) were performed for FIGURE 1 | Grassland biomass by fertilization regime. CK: non-amended; OF 3000: 3,000 kg ha −1 organic fertilizer (composted cattle manure); OF 6000: 6,000 kg ha −1 organic fertilizer; OF 9000: 9,000 kg ha −1 organic fertilizer; BOF: 3,000 kg ha −1 biofertilizer (composted cattle manure + Trichoderma inoculum); BOF 6000: 6,000 kg ha −1 biofertilizer; BOF 9000: 9,000 kg ha −1 biofertilizer. Data from each treatment were analyzed using one-way analysis of variance (ANOVA). Bars represent mean values of three replicates ± SE. Bars that do not share a letter are significantly different (p < 0.05).
Frontiers in Microbiology | www.frontiersin.org multiple comparisons. The multiple comparison correction is according to (Benjamini and Hochberg, 1995) Non-metric multidimensional scaling (NMDS) plots and a principal component analysis (PCA) were performed in R with the vegan package (Version 3.0.2, Oksanen et al., 2009). Differences in soil chemistry and bacterial and fungal communities between treatments were tested by analysis of similarities (ANOSIM) (Yan et al., 2017).
SEM is a powerful tool to examine relationships among intercorrelated variables, in which the net effects of an experimental treatment can be partitioned into direct and indirect effects. Observed variables rather than latent variables were used in the models directly. In our study, SEM was employed to test the direct and indirect pathways between soil organic matter, total N, soil chemistry, bacterial and fungal communities, Trichoderma abundance, and aboveground grassland biomass, providing a

Grassland Biomass and Crude Protein
Plots fertilized with 9,000 kg ha −1 BOF had the greatest aboveground biomass of any treatment and were significantly more productive than non-amended plots and plots fertilized with any rate of OF or BOF, except for BOF3000 treatment (Figure 1). However, there was no significant difference in plant tissue crude protein between treatments or the CK, which was similar between all amended and CK plots (mean % ranged from 6.52 ± 0.27 to 7.31 ± 0.37).

Soil Chemistry and Microbial Communities
To identify potential soil microbial ecological mechanisms underlying observed differences in grassland biomass, we focused on three representative treatments (CK, OF 9000 and BOF 9,000, hereafter referred to as CK, OF, BOF). Ethyl acetate extracts from these soil samples were analyzed by GC-MS. The GC-MS analysis revealed 37 compounds, which showed significant (p < 0.05) differences in relative abundance between treatments (Table 1), including 20 alkanes, 3 benzenes, 4 alcohols, 1 acid, 1 ether, 2 aldehydes, 3 esters, 2 amines, and 1 phenol. The relative abundance of most compounds was significantly (p < 0.05) greater in plots amended with BOF, compared to non-amended control plots, and in several cases compared to OF plots. Two-dimensional NMDS plots revealed how representative treatments related to differences in soil chemistry and microbial communities (Figure 2). For soil chemistry, pairwise contrasts indicate CK, OF, and BOF plots significantly (p < 0.05) separated by axis 1 (Figure 2A). Similarly, fungal communities showed significant (p < 0.05) variations by treatment ( Figure 2B). However, bacterial communities were relatively similar across all plots ( Figure 2C). As for soil microbial community compositions, we observed fewer bacterial genera (4) with significant (p < 0.05) differences in relative abundance by treatment compared to fungal genera (14) at the genus level (Table S1). Plots amended with BOF had greater abundance of fungal genera Archaeorhizomyces and Trichoderma but reduced abundance of Ophiosphaerella compared to OF or non-amended plots (Figure 3).
differences in plant-available P, and there were no significant difference in other soil properties. According to Pearson's correlations (Table 2), the key fungal genera abundances of  Archaeorhizomyces (p = 0.027) and Trichoderma (p = 0.035) were just positively correlated with higher plant-available P, while Ophiosphaerella (p = 0.035) was negatively correlated. As shown in Figure 4, positive correlations between grassland biomass and Archaeorhizomyces (p = 0.0068) and Trichoderma (p = 0.0048) were found, while Ophiosphaerella (p = 0.0432) was negatively correlated with aboveground grassland biomass.

SEM Pathways and Grassland Biomass
Structural equation modeling for grassland biomass was a strong fit with the data (χ 2 = 4.779, DF= 7, P = 0.687, NFI = 0.946, RFI = 0.838, IFI = 1.027, RMSEA = 0.000, Table S3). As shown in Figure 5, the explanation of the variance in grassland biomass was directly dependent on soil organic matter, Trichoderma abundance, bacterial and fungal communities, as affected by fertilization regime. Trichoderma abundance had the strongest overall effect (path coefficient = 0.879) on grassland biomass, and the effects of soil organic matter (path coefficient = 0.329), fungal community (path coefficient = 0.245), and bacterial community (path coefficient = 0.287) were also positive. However, the soil chemistry had no significant effect on grassland biomass. Fungal communities were directly mediated by total soil N (path coefficient = 1.250) and the soil chemistry (path coefficient = −1.226). Soil bacterial communities were mainly effected by the soil chemistry (path coefficient = 1.030).

DISCUSSION
Plant and microbial community structure and biodiversity are crucial for maintaining ecosystem sustainability and productivity (Bell et al., 2005;Cardinale et al., 2006). In our study, we found the application of organic fertilizer or Trichoderma biofertilizers resulted in improved aboveground grassland biomass, with optimal results from applying 9,000 kg ha −1 Trichoderma biofertilizer. Multiple mechanisms may be responsible for increasing grassland biomass. Our results indicate different fertilization regimes drove differences in soil chemistry and edaphic properties, with concomitant shifts in key soil fungal genera. This trend was most pronounced for high inputs of Trichoderma biofertilizer. Differentiation of soil chemistry between selected fertilization regimes (CK, OF, and BOF), presumably formed distinct resource FIGURE 4 | Correlations (p < 0.05) between key fungal genera (from Figure 3) and aboveground grassland biomass. niches in our grassland plots. Many of the identified chemicals were most abundant in plots amended with BOF, compared with CK or OF. Previous studies indicate some of these substances, such as pentadecane and nonadecane, have antifungal properties (Yuan et al., 2012;Raza et al., 2015). In addition, antifungal activity has been reported for the majority of benzenes, alcohols, and phenols (Raza et al., 2015). Based on our findings, we propose antifungal substances were more abundant in the BOF9000 treatment, potentially protecting plants from fungal pathogens. Future studies should focus on how soil chemical profiles influence plant pathogens and could be managed to provide greater plant protection.
Various soil chemical structures are consumed by microbes and can facilitate microbial diversity (Mwafulirwa et al., 2016). We observed a corresponding differentiation of the fungal community but not the bacterial community, based on soil chemistry, driven by fertilization regime. Martin et al. (2015) reported application of compost is a key factor explaining differences in soil microflora. Sun et al. (2015) also demonstrated the role of organic fertilizer in facilitating soil microbial community stability and diversity, which provides a foundation for grassland biomass production. Based on our correlations between microbial taxa and grassland biomass (Figure 5), we propose that Archaeorhizomyces and Trichoderma may promote plant growth, while Ophiosphaerella had the opposite effect. Trichoderma spp. are well-known for their capacity to improve plant growth and promote health in agricultural systems (Harman et al., 2004;Gravel et al., 2007;Bae et al., 2009). Examples of these mechanisms include: improving secretion plant stimulatory compounds, such as growth hormones (indole acetic acids, cytokinins, gibberellins, and zeatins; Gravel et al., 2007;Contreras-Cornejo et at., 2009), enhancing solubilization of soil nutrients (Yedidia et al., 2001;Kapri and Tewari, 2010), increasing root length and number of root hairs to absorb nutrients by exploring larger spaces of soil (Bjorkman, 2004;Samolski et al., 2012). The life cycle, ecology, and evolution of Archaeorhizomycetes remain largely unknown. However, it is understood that Archaeorhizomycetes are non-pathogenic (Rosling et al., 2011). Evidence suggests Ophiosphaerella are phytopathogens (Kaminski and Hsiang, 2008). Hciii et al. (2007) Table 2). Taken together, we conclude greater plant-available P was beneficial to aboveground plant biomass by increasing the relative abundances of Archaeorhizomyces and Trichoderma while decreasing Ophiosphaerella. Previous research has reported Trichoderma can increase plant P-uptake by increasing P-solubilization in soils (López-Bucio et al., 2015). This increased plant growth was greatest when plant-available P from composted cattle manure was provided along with Trichoderma inoculum (BOF). While our current research shows strong correlative linkages between plant, soil, and microbial factors, future research should identify if plant-available P increased plant growth directly, or via indirect effects on the abundance of Archaeorhizomyces and Trichoderma, and explore methods to better demonstrate a clear mechanistic pathway between microbial community composition, soil chemistry, and grassland biomass.
According to our SEM, soil organic matter, Total N, Trichoderma abundance, bacterial community, fungal community, and soil chemistry make a good explain for FIGURE 5 | Structural equation modeling (SEM) for grassland biomass. Numbers following variables show the percentage of its variance explained by its predictors. We utilized PCA to demonstrate variations in fungal community across different treatments, and PC1 was selected as the parameter describing the fungal community (PC1 explained 87.25% of total variation in the fungal community). Similarly, PC1 explained 44.85 and 62.77% of total variation in the bacterial community and the soil chemistry, respectively. A path coefficient is analogous to a partial correlation coefficient and describes the strength and sign of the relationship between two variables. Negative pathways are shown as blue lines, positive pathways are shown as red lines, and line thickness represents the intensity of influence. Non-significant pathways are shown in gray. Models are a good fit of our data. Model fits are given in Table S3, and significance levels are provided in Table S4. aboveground plant biomass in meadow steppe grasslands. Christian et al. (2008) demonstrated that fungal communities were most closely associated with changes in soil nutrient status, while soil pH was the best predictor of bacterial communities. This supports our results indicating total soil N had a significant influence on the fungal community and no significant effect on the bacterial community. However, Trichoderma had the greatest influence on grassland biomass. This supports the findings of Sivan et al. (1987), who demonstrated plant survival and yield can be increased, following Trichoderma inoculation in greenhouse or field settings.
Trichoderma biofertilizer (9,000 kg ha −1 ) effectively regulated soil chemistry and microbial communities, driving substantially improved aboveground plant biomass compared to organic fertilizer not containing Trichoderma. Certain soil compounds with antifungal activity may ensure individual plant fitness and increase grassland biomass. Plant-available P was beneficial to grassland biomass production, presumably by increasing the relative abundances of beneficial microbes and decreasing phytopathogenic microbes. Soil organic matter, Trichoderma abundance, and bacterial and fungal communities have direct influences on grassland biomass, while soil chemistry indirectly increases grassland biomass through alterations in bacterial and fungal communities. Among these factors, Trichoderma was the primary contributor to improve grassland biomass. Our study helps to provide a basis for grassland biomass production from the perspective of soil microbial ecology, and may ultimately improve management of grassland soils for key provisioning services. While the selected regimes in our study suggest potential underlying microbial ecological mechanisms, continuous monitoring and analyses of a gradient of fertilization rates will allow a more comprehensive mechanism understanding.

AUTHOR CONTRIBUTIONS
FZ, GY, and YZ conceived and designed the experiments. FZ and YH performed the experiments. FZ and GL analyzed the data. FZ and YZ contributed reagents, materials, and analysis tools: FZ, AC, JZ, GW, and YZ wrote the paper. All authors reviewed and contributed to the manuscript.