Effect of Trichoderma harzianum on maize rhizosphere microbiome and biocontrol of Fusarium Stalk rot

Fusarium stalk rot (FSR) caused by Fusarium graminearum (FG) significantly affects the productivity of maize grain crops. Application of agrochemicals to control the disease is harmful to environment. In this regard, use of biocontrol agent (BCA) is an alternative to agrochemicals. Although Trichoderma species are known as BCA, the selection of host-pathogen specific Trichoderma is essential for the successful field application. Hence, we screened a total of 100 Trichoderma isolates against FG, selected Trichoderma harzianum (CCTCC-RW0024) for greenhouse experiments and studied its effect on changes of maize rhizosphere microbiome and biocontrol of FSR. The strain CCTCC-RW0024 displayed high antagonistic activity (96.30%), disease reduction (86.66%), biocontrol-related enzyme and gene expression. The root colonization of the strain was confirmed by eGFP tagging and qRT-PCR analysis. Pyrosequencing revealed that exogenous inoculation of the strain in maize rhizosphere increased the plant growth promoting acidobacteria (18.4%), decreased 66% of FG, and also increased the plant growth. In addition, metabolites of this strain could interact with pathogenicity related transcriptional cofactor FgSWi6, thereby contributing to its inhibition. It is concluded that T. harzianum strain CCTCC-RW0024 is a potential BCA against FSR.

rhizosphere microbiome to control the pathogens 23 . Frequent input of the agrochemicals can reduce the phytopathogens, but it will also have a negative impact to plant associated microbiomes. In fact, beneficial microbial communities play significant role in the soil quality and fertility thereby maintaining the quality of native soil to ensure improved natural agriculture crop products. Hence the use of pathogen-specific microbial BCAs has advantage to specifically prevent the particular pathogen and to increase the beneficial rhizosphere microbiome and plant immunity.
BCAs of Trichoderma spp. have ultimate functions in promoting the plant beneficial microbial community and decreasing the pathogen attack through the specific interactions with host-pathogen. However, the effect of Trichoderma on changes of maize rhizosphere microbiome is poorly understood. According to the earlier report the transcriptional cofactor FgSWi6 is required for virulence and pathogenicity of FG 24 . Therefore, we hypothesize that inhibition of FgSWi6 can reduce the pathogenicity of FG. Hence, we screened a total of 100 Trichoderma isolates against FG by in vitro antagonistic assay. Based on screening experiments, we selected Trichoderma harzianum (CCTCC-RW0024) for greenhouse experiments to study its effect on changes of maize rhizosphere microbiome and biocontrol efficiency. The synergistic interaction between secondary metabolites and hydrolytic enzymes are crucial for BCAs 19,25,26 . Hence, we analyzed the biocontrol-related-hydrolytic enzymes, gene expression, and metabolites in the strain CCTCC-RW0024. In addition, we applied a molecular docking method to predict the inhibitory effect of T. harzianum metabolites on pathogenicity related transcriptional cofactor FgSWi6 of FG.

Results
Screening of biocontrol Trichoderma. In order to select the potent antagonist Trichoderma against FG, a total of 100 Trichoderma isolates were subjected to in vitro antagonist activity. Among the 100 isolates, the top ten potent isolates were selected for further experiments with high antagonist activity against FG such as  Table 1).
Further, the relation between hydrolytic activity and antagonist activity of Trichoderma was analyzed. The results of correlation analysis revealed that the CWDEs such as chitinase, β (1-3) glucanase, protease and cellulase displayed a positive correlation with antagonistic activity. Thus the hydrolytic activity of chitinase and β (1-3) glucanase from Trichoderma could significantly increased the mycoparasitism of 19%, and 50% respectively ( Supplementary Fig. 1).
In this assay, we analyzed the expression of biocontrol-related gene expression in selected potent Trichoderma isolates. The results indicated that biocontrol-related gene expression of chitinase (nag1, Chit33), cellulase (Thph1 and Thph2), 1,3-β-exoglucanase (exg1), aspartyl protease (papA) significantly varied between Trichoderma isolates. Among the tested isolates, the strain CCTCC-RW0024 were showed high mRNA related gene expression of biocontrol-related genes (Fig. 1). Effect of Trichoderma harzianum inoculation on physiological indices. This experiment analysed the effect of exogenous application of Trichoderma or FG treatments on physiological changes in maize seedlings. Physiological indices were measured after one month of treatment in greenhouse experiment. The average shoot length (13.31%, p < 0.05), average root length (39.59%, p < 0.05), average shoot biomass (6.38%, p < 0.05), and average root biomass (23.52%, p < 0.05) increased significantly in T4 (T. harzianum strain CCTCC-RW0024 and FG inoculated) compared to T3 FG alone treated (Fig. 2a,b). The disease reduction varied significantly between the treatments and it was high in T4 (T. harzianum strain CCTCC-RW0024 and FG inoculated) and low in T3 FG alone treated (Fig. 2c). Thus, T4 showed that T. harzianum strain CCTCC-RW0024 significantly reduced the infection of FG, as evident by 86.66% diseases reduction and increased plant growth.
Quantitative analysis of T. harzianum strain CCTCC-RW0024 in maize roots. This experiment studied the colonization of T. harzianum strain CCTCC-RW0024 in maize roots, by following two methods: (1) qRT-PCR (2) using eGFP tagged mutant of the strain by ATMT method. In the qRT-PCR method, we first designed the specific primers and confirmed the specificity by electrophoresis analysis. The results indicated that there was no visible band observed in uninoculated FG and/or strain CCTCC-RW0024 (Fig. 2d). Thus, designed primers could be accepted and used for the quantification of respective strains DNA in maize root. The calculated DNA of FG and strain CCTCC-RW0024 from qRT-PCR (Fig. 2e) showed that the DNA of strain CCTCC-RW0024 was found high in T2 (strain CCTCC-RW0024 alone treated; 2.88 ± 0.2 Log 10 DNA concentration (ng.g −1 ) of root sample) and T4 (strain CCTCC-RW0024 and FG-treated; 2.62 ± 0.26 Log 10 DNA concentration (ng.g −1 ) of root sample) whereas the FG was high in negative control T3 (FG alone treated; 2.35 ± 0.35 Log 10 DNA concentration (ng.g −1 ) of root sample) and it significantly decreased in T4 (strain CCTCC-RW0024 and FG-treated; 0.82 ± 0.12 Log 10 DNA concentration (ng.g −1 ) of root sample). In addition, the eGFP tagged strain CCTCC-RW0024 was confirmed root colonization by the observation of mycelium (Fig. 2f) and spores (Fig. 2g) under the fluorescent microscope. The microscopic observation and comparison of maize root treated and untreated eGFP tagged strain CCTCC-RW0024 clearly showed the successful colonization of the strain in maize root (Fig. 2h,i). This indicated that the strain CCTCC-RW0024 could significantly reduce the colonization of FG and its infections in maize seedlings.
Exogenous application of T. harzianum on the maize rhizosphere microbiomes. In this experiment, we analyzed the exogenous application of T. harzianum on changes of maize rhizosphere microbiomes by pyrosequencing method. The sample collections and treatments are described in materials and methods. A total of 79812 bacterial OTUs and 64304 of fungal OTUs were subjected to the classification analysis. The dominant length distributions were 693 bp for fungi and 516 bp for bacteria. The OTUs of bacteria and fungi varied significantly between the treatments. The bacterial OTUs were found higher than the fungal OTUs recorded. However, the bacterial OTUs were the maximum of 9545 OTUs in T4R (strain CCTCC-RW0024 and FG-treated) and the minimum of 7987 OTUs in T1S (uninoculated), whereas the fungi OTUs was high in T4S (8085 OTUs) and low in T2R (6647 OTUs) (Fig. 3a).
The maize rhizosphere microbiome analysis between the treatments using the best-hit revealing the distribution and sharing of the same OTUs of bacteria in root and rhizosphere soil are shown in Fig. 3b,c. A total of 142 OTUs in rhizosphere soil and 104 OTUs in maize root were found common in different treatments. The CK (uninoculated) shared the maximum of 118 OTUs with T3S (FG alone treated) followed by 106 OTUs with T4S (strain CCTCC-RW0024 and FG-treated). In the case of fungi OTUs, 40 OTUs in rhizosphere soil and 48 OTUs in maize root were common with the treatments (Fig. 3d,e).

Discussion
Maize stalk rot is ubiquitously found across the globe, in which the causal organism Fusarium graminearum is able to infect root and basic stem throughout plant growth with typical symptoms at late stage of the growth at grain filling of maize development. So far, there is no good measure to control the soil-borne disease. In recent years, we established Trichoderma as potent BCAs to treat maize seeds or rhizosphere soil in farming fields, which has resulted in a significant decrease of disease by 60% or above (data not shown). Trichoderma spp., are known to reduce the plant pathogenic attack and also increase the plant growth and yield through the enhancement of growth hormones, soil fertility, and increment of plant beneficial microbiome [27][28][29] . The biocontrol performance of Trichoderma strains in the farming field varies based on environments impact, host-pathogen specificity and stability. Therefore, isolation and screening of Trichoderma species against FG from different ecosystems according to the host-pathogen specificity could provide enhanced insight of BCAs. In this study, we screened a total of 100 Trichoderma isolates against the FG, and reported a potent strain T. harzianum CCTCC-RW0024 with significant antagonistic activity (96.30%), disease reduction (86.66%), biocontrol-related hydrolytic enzyme and gene expression.
Also, we found that this strain significantly increased the maize rhizosphere beneficial microbial community. The pyrosequencing indicated that exogenous inoculation of the strain CCTCC-RW0024 in maize rhizosphere could increase plant growth promoting acidobacteria (18.4%), decreased 66% of FG and increased the plant growth. This strain could also increase the plant beneficial microbiome in maize rhizosphere as evident by qRT-PCR and electrophoresis assay of FG. Similarly, several researchers have reported that acidobacteria act as plant growth promoting bacteria 30,31 .
Colonization of the Trichoderma strains in maize root playsan important role in the function of mycoparasitism process 12 . This was confirmed through qRT-PCR quantification and microscopic observation of eGFP tagged strain T. harzianum CTCC -RW0024 in maize root. Moreover qRT-PCR results indicated that the colonization of strain CCTCC-RW0024 significantly reduced the FG in maize root (Fig. 2e). Similarly, Trichoderma strains are known to prevent colonization of FG in maize roots 1, 2 . The results also revealed that exogenous inoculation of T. harzianum strain CCTCC-RW0024 could significantly increase the plant growth, plant biomass and reduce the pathogen attack.Similar results are also found in the work of Akladious and Abbas 32 .
Antagonistic activity of the Trichoderma strains is generally related to the hydrolytic enzyme activity especially chitinase (chitin) and β (1,3) glucanase (pachyman). These enzymes are closely related to mycoparasitism, and on the other hand, they are contributing to the host induced systemic resistance (ISR) to pathogen infection 4, 33, 34 . In our case, CWDEs from Trichoderma CCTCC-RW0024 played multiple functions in the biocontrol process against stalk rot 35,36 (Supplementary Table 2, Fig. 1). We also found that there was a positive correlation of 19%, and 50% of the antagonistic activity and enzyme activity of chitinase and β (1,3) glucanase respectively ( Supplementary Fig. 1). This means that biocontrol activity of Trichoderma strain depends on the combined action of multiple enzymes.

Conclusion
This work reports T. harzianum strain CCTCC-RW0024 as a potent biocontrol agent against Fusarium graminearum a causal agent of maize stalk rot. In addition, this work indicated that inoculation of T. harzianum strain CCTCC-RW0024 in infected maize field could control stalk rot through different mechanisms or composite mechanisms including mycoparasitism, ISR, the increment of the maize rhizosphere microbiome and soil fertility. This work emphasized to study the genetic, transcriptomes and specific characterization of SMs in T. harzianum strain CCTCC-RW0024 responsible for the interaction of maize root. Therefore, further molecular study  and SMs characterization will help to understand the molecular mechanisms through which T. harzianum acts as biocontrol agents.

Materials and Methods
Fungal isolates, Medium, and culture preservation. A total of 100 Trichoderma isolates were tested in this study for control of stalk rot fungal pathogen Fusarium graminearum (FG). Trichoderma isolates used in this were previously isolated from the non-agricultural ecosystem in China 39 . The culture of F. graminearum was received from Prof. Dr. Jie Chen, Shanghai Jiao Tong University, P. R. China.
In vitro assay. In this assay, a total of 100 Trichoderma isolates were selected to test the antagonistic activity according to the method of Dennis and Webster 40 . In detail, the mycelial disk (5 mm diameter) of 7 days old growing edge of Trichoderma and FG were placed on PDA Petri disk (Size-90 × 15 mm) at an equal distance of opposite direction. The control was maintained without Trichoderma disk. The incubation and the method of incubation rate calculation were described in our previous study 19 . RNA was isolated from the fungal biomass and treated with DNase I according to manufacturers instructions (RNAprep Pure Kit, TIANGEN Biotech (Beijing) Co., Ltd.). It was then reverse transcribed to cDNA according to manufacturer instructions (PrimeScript TM II 1 st Strand cDNA synthesis kit, Takara, Clontech, Japan). cDNAs were diluted 1:10 and analysed for the CWDEs such as chitinase (nag1), cellulase (Thph1 & Thph2), protease (papA) and glucanase (exg1) gene expression by using SuperReal PreMix Plus (SYBR Green; Tiangen Biotech Co. Ltd, China) by an FTC-3000 Real-Time quantitative PCR System (Funglyn Biotech, Canada). The actin gene (act) was used as control (housekeeping gene) 44 . Primer for the Thph1 and Thph2 was designed in this study and other primer sequences were retrieved from Steindorff et al. 45 .

Biocontrol
Antagonistic activity in planta. Maize (Inbred line Huangzao 4) seeds were surface sterilized according to the method of Morán-Diez et al. 46 . The seeds were then allowed to germinate on sterile wet filter paper at 25 °C for 72 h. For the greenhouse experiments, for the in planta antagonistic assay, the native soil was collected from local agriculture field. In order to minimize the weeds, nematodes, insects, microbiome in soil, the native soil was subjected to solarization followed by sterilization at 180 °C for 6 h. The greenhouse experiments were carried out in the soil with pH 7.1, and organic matter (0.7%). We selected T. harzianum CCTCC-RW0024 for greenhouse experiment based on the screening. Spore suspension of the strain CCTCC-RW0024 was prepared aseptically for greenhouse experiment according to the method of Cappuccino and Sherman 30 . Five ml of CCTCC-RW0024 suspension (3.8 × 10 7 CFU/ml) was inoculated in the soil. For pathogen inoculation, 1.5 g of FG biomass was inoculated in one kg of soil. The greenhouse experiment was conducted inside the greenhouse with a temperature of 27 ± 2 °C and photosynthetically active radiation of 750 ± 75_mol.m −2 .s −1 . Two greenhouse experiments were performed with the objectives (1) to study the in planta antagonistic activity. For this experiment, we used sterile soil, and (2) to study the impact of the exogenous application of T. harzianum CCTCC-RW0024 on the maize rhizosphere microbiome, for this experiment we used unsterilized native agriculture soil.
A total of four treatments were applied for the greenhouse experiment as follows: CK1-Natural farming soil as such collected from the farming field. For T. harzianum strain CCTCC-RW0024 colonization analysis T 1 -Natural farming soil not inoculated with any microbes and amended with 10% vermiculite, T 2 -Natural farming soil inoculated with T. harzianum strain CCTCC-RW0024 (T72) and amended with 10% vermiculite, T 3 -Natural farming soil inoculated with FG and amended with 10% Vermiculate, T 4 -Natural farming soil inoculated T. harzianum strain CCTCC-RW0024 (T72) and FG and amended with 10% Vermiculate. The pre-germinated seedlings were planted in the pot after a week of microbial inoculation. One month later, shoot length, root length, shoot biomass, and root biomass were measured in the maize seedlings. Fusarium stalk rot disease reduction (%) in maize by T. harzianum CCTCC-RW0024 treatments was evaluated at an adult stage after the three months of treatment. Diseases reduction was assessed using the formula described by Saravanakumar et al. 19 . These experiments were conducted as three replicate per treatments in a randomized trial to study the effect of each treatment on growth of maize.

RT-qPCR detection of T. harzianum and FG colonization.
In order to determine the fungal colonization in the maize root, the ITS copies of T. harzianum CCTCC-RW0024 and FG were determined using RT-qPCR by designing specific primers for T. harzianum CCTCC-RW0024: ITS spacer TH-F (5′-GCATTTCGCTGCGTTCTTCA-3′), ITS spacer TH-R (5′-TAATCTGAGCCTTCTCGGCG-3′), F G : I T S s p a c e r F G -F ( 5 ′ -G AG G G T T G A A AT G AC G C T C G -3 ′ ) a n d I T S s p a c e r F G -R (5′-ACGGATCTCTTGGTTCTGGC-3′) colonization was determined in all the four treatments.
For the RT-qPCR assay, three plant roots were randomly collected from each treatment. Maize roots samples were prepared according to the methods described by Saravanakumar et al. 19 . The root DNA was extracted from 100 mg of fresh maize root by using a DNA extraction kit (Plant Genomic DNA Kit, TIANGEN Biotech (Beijing) Co., Ltd.). The number of internal transcribed spacer (ITS) copies of Trichoderma and FG were determined by using Super Real PreMix Plus (SYBR Green, TIANGEN Biotech (Beijing) Co., Ltd.) by an FTC-3000 Real-Time quantitative PCR System (Funglyn Biotech, Canada).
Scientific RepoRts | 7: 1771 | DOI:10.1038/s41598-017-01680-w Generation of eGFP-Tagged T. harzianum. The eGFP expressing cassette sequence was amplified from the plasmid pCPXHY1eGFP with the primers Pgpd-F/Tgpd-R 47 . The eGFP expression of T. harzianum CCTCC-RW0024 was generated by ATMT method using the modified plasmid pCAMBIA1300. Resistant colonies were observed on PDA plates supplemented with 200 µg/mL hygromycin. The presence of green fluorescence and the insertion of eGFP expression cassette were confirmed by fluorescent microscopic observation of T. harzianum CCTCC-RW0024.

Analysis of microbiomes by pyrosequencing.
In order to analyse the microbial populations the unsterilized native soil were used, a total of four soil, four maize root and one natural farming soil samples such as CK-unsterilized natural farming field soil, T1S-uninoculated soil, T2S-Soil inoculated with T. harzianum strain CCTCC-RW0024, T3S-Soli inoculated with FG, T4S-Soil inoculated with T. harzianum strain CCTCC-RW0024 and FG, T1R Maize root samples collected from uninoculated soil pot, T2R maize root collected from soil inoculated with T. harzianum strain CCTCC-RW0024, T3R-Maize root samples collected from inoculated with FG pot, T4R-Maize root samples collected from inoculated with T. harzianum strain CCTCC-RW0024 and FG were subjected to the pyrosequencing experiments. Genomic DNA was extracted from the soil, maize root samples using the TIANamp sediment and plant DNA kit according to manufacturer's instructions (Tiangen, China). After extraction, the purity of the DNA was tested using a UV spectrophotometer followed by 0.8% agarose gel electrophoresis at a voltage of 120 V for 20 min. The internal transcribed spacer (ITS) regions were amplified for fungal ITS1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and 5.8 sRNA gene A2R (5′-CTGCGTTCTTCATCGAT-3′) and bacterial 16S rRNA gene: F343 (5′-TACGGRAGGCAGCAG-3′) and R803 (5′-CTACCAGGGTATCTAATCC-3′) 48 . The PCR reaction was conducted in 20 ng/µl of reaction mixture which consisted of 8.75 µl of ultra pure H 2 O, 5 µl of 5x Q5 Buffer, 5 µl of 5x GC Enhancer, 2 µl of dNTP (2.5 mM), 2 µl of template DNA (2 ng/ µl), 1 µl each forward and reverse primer (10 µM), 0.25 µl of Q5 DNA polymerase. The PCR cycling conditions were set as 4 min at 98 °C, 27 cycles (98 °C for the 30 s, 47.6 °C for 45 s, 72 °C for 1 min), then the final extension at 72 °C for 5 min, and finally the experiment was halted at 10 °C. The PCR products were purified using AMpure Beads followed by the PicoGreen dsDNA assay kit used for the quantification of DNA. Finally, the mixture was pyrosequenced by using Roche 454 GS FLX (Shanghai Personalbio Co., Ltd., China). The sequence data was analyzed by using Qiime (version 1.7.0, http://qiime.org/) followed by the mother (version 1.31.2, http://www.mothur.org/) and further the data was analyzed according to Sun et al. 49 . In brief, the below sequences score of 25 and 200 bp length were trimmed and binned into operational taxonomic units (OTUs) using a 97% of the threshold for the bioinformatics and subsequent analysis 49 . Extraction of secondary metabolites. T. harzianum CCTCC-RW0024 was subjected to extraction of fungal metabolites according to the methods described by Saravanakumar et al. 19 . In brief, the culture of CCTCC-RW0024 (3.5 × 10 7 spores/ml) was inoculated into 20 liters of a production medium with the pH 7.2 and incubated for 31 days at 28 °C at 180 rpm of shaking. After the incubation, the microbial culture was subjected to ethyl acetate extraction and the crude extract was concentrated using a rotary evaporator. Finally, the concentrated extracts were dissolved in 90% methanol for further partial purification. A total of 10 fractions were collected from the column chromatography and subjected to antifungal activity assay against plant pathogen FG at different concentrations (1-100 µl. 250 ml −1 of PDA medium). The selected active fractions were partially studied by the gas chromatography-mass spectrometer (GC-MS; AutoSystem XL GC/TurboMass MS).
Preparation of transcriptional cofactor FgSWi6. In order to study the inhibitory effect of Trichoderma metabolites on pathogenicity related transcriptional cofactor FgSWi6 of FG (Liu et al. 24 ). An in silico molecular docking method was applied 19 . The FASTA format of an amino acid sequence of FgSWi6 (sequence length of 807 residues and molecular weight of 87663.2) was retrieved from protein database of National centre for Biotechnology information (NCBI). We used the ClustalW and MODELLER v9.12 for the sequence alignment for FgSWi6 with the template protein. The 3D structure FgSWi6 was predicted using template amino acids sequence by Swiss-Model (https://swissmodel.expasy.org/). Of the three models, the one with the lowest score was taken for energy minimization using the CHARMm force field in Discovery Studio v4.1. The modeled 3D structure of FgSWi6 was used for molecular docking studies.

Molecular docking analysis. Fungal metabolites were partially identified based on GC-MS analysis.
Further the compound structures were predicted by using the ChemSketch software (http://www.acdlabs.comfrom) with the assistance of NCBI-PubChem database (http://www.ncbi.nlm.nih.gov/pccompound). The predicted structure of compound molecule was saved as MOL file for docking study 19 . The interaction between the Trichoderma-derived metabolites and target transcriptional cofactor FgSWi6 was studied by using ArgusLab 4.0.1 and Discover Studio Version 4.0 (Accelry's Software Inc. the USA). The mechanism of ligand placement was based on binding site position, therefore, the fitting points were added to hydrogen bonding groups on the target protein and Trichoderma-derived molecule (ligand). Scoring function was implemented in docking programme by making various assumptions and implications in the evaluation of model complexes, which includes terms of hydrogen bonds employed by Discovery Studio to rank the docked bases and to assess the binding site and the number of present rotatable bonds.

Statistical analysis.
All experiments were repeated three times. The significance between variables among Trichoderma species in enzymes experiments and treatments of greenhouse experiments, a suite of statistical analysis (SPSS 11.5) and MS Excel-2007 was used to find the mean and standard error. ANOVA (one-way classifications) with Duncan Post hoc multiple comparisons was applied. The heat map for the mycoparasitism related genes was drawn by using HemI: A Toolkit for Illustrating Heatmaps 50 .