Trichoderma fungi were isolated from 13 soil samples (each sample was approximately 100 g) of white birch rhizosphere soil (at a depth of 5–15 cm). Of these soil samples, 10 were collected from planted forest at the forest farm of Northeast Forestry University and three were collected from the natural forest of Maoer Mountain (126.63°E, 45.72°N). From these soil samples, five 5 g subsamples of each sample were dissolved in 100 mL of sterile water, shaken well using a muddler, and made into stock solution, which was diluted to four concentrations (1/10, 1/100, 1/1000, and 1/1000). Each concentration of the solution (200 μL) was plated onto potato dextrose agar medium (PDA medium) in a 9 cm Petri dish and cultured at 28°C for 24–96 h. During this period, the Petri dishes were observed every 24 h, and the Trichoderma colonies were transferred to a new PDA medium for purification culture (Zhou et al., 2019).

The purified Trichoderma were inoculated on a PDA plate medium and cultured in the dark for 7 days at 28°C to observe their macroscopic morphology; images of the colonies were obtained. Some PDA medium blocks (their length, width, and height were approximately 0.5 cm × 0.5 cm × 0.1 cm) were prepared using a sterile blade and placed on a glass slide under aseptic conditions. Then, the Trichoderma spores were inoculated on the PDA medium block and the aseptic coverslip was covered to make hyphae grow on the coverslip. The Trichoderma fungi were cultured in the dark at 28°C for 48–72 h, and the relative humidity was 60%. The hyphae were stained with a blue dye (biohao Biotechnology Co., Ltd., Beijing, China, Catalog No. C0710) to observe and record their morphology and that of their spores under a microscope (Leica DM750, Wetzlar, Germany). Morphological identification relied on the descriptions found in previous research (Yang, 2009, in Chinese).

One isolate in each species identified by morphology was randomly chosen for molecular identification. The purified Trichoderma was inoculated in potato dextrose (PD) liquid medium and shaken at 28°C at 180 rpm for 48 h. The mycelia were filtered and collected using eight layers of gauze for DNA extraction using a DNA extraction kit (OMEGA Biotek, Beijing, China). Then, the extracted DNA were used as a template to amplify the ITS (internal transcribed spacer) region and tef1-α (translation elongation factor-1α) region; the primers were designed with reference to previous studies (Dou et al., 2019), The PCR products were purified using a gel extraction kit (Promega, Madison, WI, USA, Kit No. A9281) and subjected to direct automated sequencing using fluorescent terminators using an ABI 377 Prism Sequencer (Sangon Biotech, Shanghai, China). The sequencing results were compared and identified using the website of the Center for Culture Collection of Trichoderma (CCTC), Shanghai Jiaotong University.¹ Other recommended reference Trichoderma gene sequences were selected from the database (Dou et al., 2019), and a phylogenetic tree was constructed using the maximum likelihood (ML) method, with 1,000 bootstrap replications in the MEGA 7.0 (Kumar et al., 2016) package. After identification, the sequences were submitted to Genbank.² The Trichoderma isolates that underwent molecular identification were subjected to further analyses.

To investigate the tolerance of the Trichoderma isolates to a harsh environment (saline alkali soil or barren soil), a culture medium simulating a harsh environment was established. Salt stress, alkali stress, and nutrition stress experiments were created using a diluted PDA medium (diluted to 1/2, 1/4, 3/4, 1/8, 1/16, or zero, only made up of agar and water), a concentrated PDA medium (concentrated to 2-fold or 4-fold), a PDA medium to which was added NaCl (to generate final concentrations of 2, 4, 6, 8, and 10%, w/v, g/mL), or a PDA medium to which was added NaHCO₃ (to generate final concentrations of 0.1, 0.4, 0.7, 1, and 1.3%, w/v, g/mL). Then, 3 μL of spore suspension (10⁶ spores/mL) was pipetted out and inoculated into the salt stress, alkali stress, and nutrition stress media, after which the Petri dishes were left to stand for 5–10 min and then sealed using parawax film. The Trichoderma were cultured at 28°C and the colony radii was measured using a ruler at 48 h.

In order to analyse the antagonism of the Trichoderma isolates, dual culture was performed between Trichoderma and phytopathogens on a PDA medium. Purified cultured Trichoderma and Cytospora chrysosperma, Botrytis cinerea, A. alternata, and Fusarium oxysporum were prepared as inocula. A 5 mm diameter mycelial disc of each Trichoderma isolate was added to the periphery of a new aseptic Petri dish; opposite this placement, a 5 mm diameter mycelial disc of the phytopathogen was positioned. Each dual culture was set up with three replicates. The plates were cultured for 12 days, and a modified resistance evaluation was performed based on a previously described protocol (Popiel et al., 2008), which used the following scoring system: +8 meant the antagonist had fully grown on the plate, completely covering the plate and the phytopathogen; +6 meant the antagonist occupied 85% of the plate’s surface area; +4 meant the antagonist occupied 70% of the plate’s surface area; and 0 meant the antagonist occupied 50% of the plate’s surface area.

A spore suspension of T. atroviride was prepared to a concentration of 10⁷ spores/mL. Twenty seeds of Gynura cusimbua (bought from the Taobao online shopping platform³) were planted in each pot (size: L × W × H = 40 cm × 28 cm × 6 cm); for the treatment group, 800 mL of spore suspension was evenly poured into the soil before planting, while for the control group, 800 mL of water was evenly poured into the soil before planting. Three replicates were performed. The seedlings were cultivated under a 16 h light/8 h dark photoperiod (under a daylight lamp) and 40% relative humidity and watered every 10 days with 500 mL of additional water. The germination rates and germination times of the plants were recorded. At 30 days after seed sowing, the stem height and stem diameter were recorded, and the leaves of three randomly chosen seedlings were harvested and stored at −80°C for malondialdehyde (MDA), catalase (CAT) activity, soluble sugar content, and soluble protein content analyses (Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China, Kit A003-1-2, Kit A00-1-1, Kit A145-1-1 and Kit A045-2-2).

Then, the 6–10 randomly chosen leaves of other plants in the control group and treatment group were pierced (4 times on each leaf) and inoculated with A. alternata, marked as C + A (control + A. alternata) and T + A (treatment + A. alternata). After 12 days, the morbidity was observed, and the MDA, CAT activity, soluble sugar content, and soluble protein content were also analysed.

Approximately 200 g of soil samples in the control group and treatment group were collected, marked as the P group (the soil in which G. cusimbua was planted) and T + P group (the soil watered with the spore suspension of T. atroviride in which G. cusimbua was planted), and stored at −20°C for further analysis.

Six pots (size: L × W × H = 40 cm × 28 cm × 6 cm) containing the same soil as that in the experiment above were prepared; for three of these, 800 mL of spore suspension of T. atroviride was evenly poured into the soil, while for the other three, 800 mL of water was evenly poured. The pots were placed under a 16 h light/8 h dark photoperiod (under a daylight lamp) and 40% relative humidity and watered every 10 d with 500 mL of additional water for 30 days (three times). Then, the soil samples (approximately 200 g) were collected, marked as the T group (the soil watered with the spore suspension of T. atroviride) and the CK group (the soil watered with water) and stored at −20°C for further analysis. The available nitrogen and available phosphorus of the soil in the P, T + P, T, and CK groups were analysed according to previous research (Bardgett et al., 2007; Zhou et al., 2020).

All experimental data were subjected to an analysis of variance (ANOVA) using SPSS software v19.0 (IBM Corp., Armonk, NY, USA). The statistical significance of the differences between the means was determined using Duncan’s multiple-range tests and an independent-sample T test (p < 0.05).

In total, 37 Trichoderma isolates were obtained from 13 soil samples, and 5 species were identified based on morphological comparison (Figure 1): T. pleuroticola (3 isolates), T. virens (2 isolates), T. hamatum (8 isolates), T. atroviride (21 isolates), and T. asperellum (3 isolates). One isolate was randomly selected from each Trichoderma species for further molecular identification. The ITS regions and tef1-α regions of these five Trichoderma isolates were amplified and sequenced. The sequences were by consulting the Center for Culture Collection of Trichoderma (CCTC), Shanghai Jiaotong University (see footnote 1). The alignment results show that most morphological identification was accurate, except for the T. asperellum sample, which was identified as T. asperelloides based on molecular characteristics. Phylogenetic trees were also constructed using the reference sequences recommended by the CCTC database (Figure 2); the results accompany the alignment results. Based on the above results, five species of Trichoderma were obtained: T. pleuroticola, T. virens, T. hamatum, T. atroviride, and T. asperelloides. The sequenced ITS and tef1-α sequences of the five Trichoderma isolates were submitted to Genbank (see footnote 2), and the accession numbers of these ITS sequences are MK377309 (T. pleuroticola), MK377310 (T. virens), MK377311 (T. hamatum), MK377312 (T. atroviride), and MK377313 (T. asperelloides). The accession numbers of the tef1-α sequences are OL547398 (T. pleuroticola), OL547399 (T. virens), OL547400 (T. hamatum), OL547401 (T. atroviride), and OL547402 (T. asperelloides). Among these isolated Trichoderma, the number of isolates of T. atroviride was the highest (21 isolates), accounting for 57% of the total isolated Trichoderma. T. hamatum followed with 8 isolates, accounting for 22% of the total. T. asperelloides and T. pleuroticola both had three isolates, accounting for 8% of the total. There were the least T. virens isolates, with two strains, accounting for 5% of the total (Figure 3).

In this study, all five Trichoderma candidate strains for which molecular identification was performed could grow in the five concentrations of the salt stress medium, among which T. asperelloides grew better and had stronger salt stress adaptability, followed by T. hamatum, which also had relatively good salt stress tolerance. In the alkaline stress medium, T. atroviride could not grow when the alkali concentration reached 0.7% (weight/volume, g/mL), while T. virens grew slowly when the alkali concentration reached 1.3% (weight/volume, g/mL). T. virens and T. pleuroticola have better adaptability to alkali stress. In the media with different nutrient concentrations, T. pleuroticola, T. virens, and T. asperelloides grew better (Figure 4).

The confrontation assay shows that the five Trichoderma isolates spread in the Petri dish rapidly reacted to and had varying degrees of antagonistic responses against F. oxysporum, A. alternata, B. cinerea, and C. chrysosperma (Figure 5). Among them, T. asperelloides could not only inhibit the growth of the four pathogens, but also parasitised the tested pathogens, occupying the entire plate; thus, it had the highest evaluation score (Figure 5B). With the exception of F. oxysporum, the other three pathogens were also parasitised by T. atroviride, T. pleuroticola, and T. virens (Figure 5A); thus, those species also have relatively high evaluation scores. T. hamatum could only parasitise C. chrysosperma but could not parasitise the other three pathogens (Figure 5A).

In the treatment group (T), the seed germination rate of the treatment group was 90% and the average germination time was 102.3 h, while in the control group, the germination rate was 80% and the average germination time was 105.2 h; this difference was not significant (Table 1). Thirty days later, we found that the seedlings in the treatment group had better growth than those in the control group (Figure 6). The average plant height of the plants in the treatment group was 4.60 cm, significantly higher than that (4.12 cm) of the control group (p < 0.05). Similarly, the average stem diameter of the plants in the treatment group (5.98 mm) was also significantly higher than that (5.04 mm) of the control group (p < 0.05, Table 1). After inoculation with A. alternata, the height and stem diameter of plants did not significantly change in the treatment group (T + A) and control group (C + A).

Further, some physiological indices were measured. For the treatment group, the MDA was 31.709 μmol/100 g, significantly higher than the MDA of the control group (16.13 μmol/100 g, p < 0.05). The soluble protein for the treatment group was 1,555.458 μg/g, significantly higher than that of the control group (1,161.634 μg/g); the soluble sugar of the treatment group was 1.5362 mg/g, significantly higher than that of the control group (1.3238 mg/g, p < 0.05). The catalase activity showed similar results, with the activity in the treatment group (1,053.63 nmol/min g) significantly higher than that in the control group (256.95 nmol/min g, p < 0.05).

After inoculation with A. alternata, the MDA values in both T + A and C + A were increased, but the MDA in C + A increased more dramatically (from 16.13 μmol/100 g to 39.077 μmol/100 g) and was significantly higher than the MDA of the plants in T + A (35.797 μmol/100 g, p < 0.05). The soluble protein of the plants increased in both groups, its value in T + A (1,694.710 μg/g) being still significantly higher than that in C + A (1,397.893 μg/g, p < 0.05). The soluble sugar of the plants also increased in both groups, and its value in T + A (1.7677 mg/g) was also still significantly higher than that in C + A (1.664 mg/g, p < 0.05). The CAT activity in plants of both groups was increased, and its value in T + A (1,773.78 nmol/min g) was significantly higher than that in C + A (735.70 nmol/min g, p < 0.05). Meanwhile, the plants in the T + A group grew better and had less or smaller disease spots compared to the plants in C + A group, which indicates that T. atroviride improves the disease resistance of G. cusimbua.

The contents of available nitrogen and available phosphorus in the four groups were measured (Figure 7). The available nitrogen in the P group, T + P group, T group, and CK group was 38.7 mg/kg, 41.33 mg/kg, 51.43 mg/kg, and 47.5 mg/kg, respectively (Figure 7A). The available phosphorus in the P group, T + P group, T group, and CK group was 1.9 mg/kg, 2.33 mg/kg, 5.2 mg/kg and 3.4 mg/kg, respectively (Figure 7B). The results show that the use of T. atroviride significantly improves the content of available nitrogen and available phosphorus in soil. Meanwhile, compared with the soil in which G. cusimbua was not planted, the soil in which G. cusimbua was planted had significantly lower available nitrogen content (p < 0.05, Figure 7A) and lower available phosphorus (Figure 7B). Furthermore, the decrease in available nitrogen from T to T + P (10.1 mg/kg) was more than the decrease from CK to P (8.8 mg/kg); the decrease in available phosphorus from T to T + P (2.87 mg/kg) was more than the decrease from CK to P (1.5 mg/kg), which indirectly shows that T. atroviride promotes the absorption of available nitrogen and available phosphorus in G. cusimbua.