The highly aggressive Fusarium oxysporum f. sp. cubense tropical race 4 (FocTR4) 14013 strain used in this study was isolated from the diseased roots and stems of the Cavendish banana (Musa spp. AAA group) and identified (Liu et al., 2023). This pathogenic strain was kept at 4–10°C and −80°C in both the Culture Collection of Innovative Institute for Plant Health, Zhongkai University of Agriculture and Engineering, Guangdong Province, China and Guangdong Microbial Culture Collection Centre (GDMCC, 61934).

Cavendish banana plants (Musa spp. AAA group) were used in this study. Healthy banana tissue culture seedlings were transplanted into 90 mm diameter pots using a sterile peat-coconut coir mix. The seedlings were randomly distributed in a culture chamber at 28°C with a 16 h light/8 h dark photoperiod. After around 8 weeks, plants with five leaves and a healthy root system were chosen for Foc inoculation, which was undertaken as described by Dong et al. (2020), with minor modifications. Plant roots were soaked for 30 min in a 2 L spore suspension containing 10⁷ conidia/mL of the 14,013 FocTR4 strain and then planted in pots.

Five dendrobe plants were collected from a farm in Panyu, Guangzhou, China (22°93’N, 113°38′E). All samples were stored at 4°C in the refrigerator until endophyte isolation. Isolation was performed within 48 h of collection according to the method described by Dong et al. (2021). Healthy dendrobe roots were initially washed with running tap water for several minutes and then with sterile water for three times. The roots were cut into about 3 mm-long pieces and each piece was then surface sterilized by being dipped into 75% ethanol for 30 s and 2.5% sodium hypochlorite (NaClO) for 20–60 s, before being rinsed three times with sterilized water. Then the cuttings were dried on sterilized filter paper and placed on the surface of Trichoderma selective medium (TSM: MgSO₄·7H₂O 0.2 g; K₂HPO₄ 0.9 g; KC1 0.15 g; NH₄NO₃ 1.0 g; glucose 3.0 g; chloramphenicol 0.25 g; rose-bengal 0.15 g; agar 20 g with 1 L water) plates. The spread plate method was used to test whether the sterilized water of the last rinse contaminated by other fungi or not (Dong et al., 2021). Finally, the plates were incubated at 28°C for 2–3 d. Colonies growing on TSM were then transferred to the water agar (WA) plates for single spore isolation. After that, the isolates were cultured at potato dextrose agar (PDA) plates. Single spore isolates were stored at 4°C until further use in this study.

Total genomic DNA was extracted from mycelia grown on PDA and incubated for approximately 5–7 d at 28°C using the modified cetyltrimethylammonium bromide (CTAB) method. Primers ITS5 (5’-GGAAGTAAAAGTCGTAACAAGG-3′) and ITS4 (5’-TCCTCCGCTTATTGATATGC-3′) were used to amplify the rDNA ITS regions (White et al., 1990). Primers EF1-728F (5’-CATCGAGAAGTTCGAGAAGG-3′, Carbone et al., 1999), and EF1-986R (5’-TACTTGAAGGAACCCTTACC-3′, Jaklitsch et al., 2005) were used to amplify the translation elongation factor 1 (TEF1) gene. The amplicons were visualized using 1.5% agarose electrophoresis gel. Positive amplicons were sequenced by Tianyi Huiyuan Biotechnology Co., Ltd., Guangzhou, China.

The sequences obtained in the present study were checked by BioEdit v 7.25 (Hall, 2006) and blast in the NCBI¹ by using the blastn. For the phylogenetic analysis, reference sequences for Trichoderma species and related taxa were obtained from the NCBI GenBank database. Each locus was aligned using the MAFFT² (Katoh et al., 2019). The alignment results were trimmed using the Trimal tool and then combined by the PhyloSuite (v1.2.1). Phylogenetic analyses were runned based on maximum likelihood (ML) in RAxML v. 8.2.12 (Stamatakis, 2014), maximum parsimony (MP) in PAUP (v.4.0b10) (Swofford, 2002) and Bayesian analysis (BI) in MrBayes (v. 3.0b4) (Ronquist and Huelsenbeck, 2003).

The dual culture technique was used to screen isolates with antagonistic activity against FocTR4 14,013 strain and other pathogens, such as Alternaria alternata, Calonectria ilicicola (Chen et al., 2022), Colletotrichum gloeosporioides, F. oxysporum from Morinda officinalis (Dong et al., 2019), Helminthosporium maydis, Lasiodiplodia pseudotheobromae (Chen et al., 2020), and Stagonosporopsis pogostemonis (Luo et al., 2022). All of the strains were cultured by PDA plates at 28°C. A 5 mm in diameter disc from the margin of a 3 days-old Trichoderma colony was placed on one side of a PDA plate (8.5 cm diameter) together with another 5 mm disc from freshly cultured mycelia of pathogens placed on the opposite side with a 5 cm diameter distance. Each treatment was replicated five times.

After an incubation period of 7 days at 28°C, the growth inhibition of the pathogen colony was calculated in comparison to the control, according to the following formula: Ig (%) =  C g − T g C g − 0.5  × 100, where Ig (%) is the percentage inhibition of radial mycelial growth, Cg is the radius length (mm) of the pathogen colony in the control, and Tg is the radius length (mm) in the dual-culture. Mean percentage ± standard error of the mean (n = 5) was calculated.

The banana tissue culture seedlings were bought from company and then cultured with sterile matrix soil. Four leaf stage healthy banana seedlings with fresh weights of approximately 6 g were selected for the experiments. Trichoderma isolate Tk905 was cultured on PD for 7–10 days with the 180 rpm/min at 28°C. The mixture were then centrifuged at 10000 rpm/min at 4°C to collect the mycelium and spores. For each seedling, a total of 10 g mycelium and spores were collected from cultures, washed with sterile water three times and transferred in a container with 200 mL of sterile water, then mixed with soil. Finally, an additional 250 mL of H₂O was added in both with and without Trichoderma soil. This mixture was maintained for 5 days, and subsequently was added H₂O daily to the soil. Then after 5 days, the banana seedlings were transplanted to the soils with the fungal suspension (T+ treatment, with Trichoderma strain Tk905) or with water in control plants (T−, without Trichoderma strain Tk905). Each pot was watered and cultured under natural conditions at approximately 28–32°C temperature. Ten seedlings were used for each treatment. All experiments were performed in triplicate.

After 26 d of growth, root length, root weight, crown weight, plant height, ball diameter, crown dry weight, and root dry weight were measured. Each pot was submerged in water for 30 min at room temperature, and the plants were lightly shaken until the substrate was completely removed to avoid root loss or damage during harvesting. The roots were then rinsed with tap water until all visible sand and soil particles were removed.

The root system was inoculated in T+ and control plants (T−) as described in the previous section. 1 g of Cavendish banana leaves tissue and roots from various treatments (T+, T−) in different time sampling (0, 12, 24, 48, and 72 h after Trichoderma strain Tk905 inoculation) were homogenized using 5 mL ice-cold 100 mM phosphate buffer (pH 6.4) containing 0.5 polyvinyl pyrrolidone (PVP) and centrifuged at 6000 rpm for 50 min at 4°C. The activity of polyphenol oxidase (PPO) was determined by NBT reduction method (Anderson and Morris, 2011). The peroxidase (POD) was determined by guaiacol method (Ciou et al., 2011). For catalase (CAT) extraction, a 0.05 M phosphate buffer (pH 7.0) was used following Naeem-Abadi and Keshavarzi (2016). The phenylalanine ammonia-lyase (PAL) was detected according to Martinez-Tellez et al. (1998). All enzymatic activity data are presented as the mean of three individual samples at each time point. Total soluble protein content was determined according to Bradford (1976), using bovine serum albumin as a standard.

The compatibility of T. koningiopsis Tk905 with commercially available fungicides against FocTR4 14,013 was determined. Seven fungicides containing mefentrifluconazole + pyraclostrobin, pyraclostrobin + fluxapyroxad, fluopyram + trifloxystrobin, propiconazole + azoxystrobin, difenoconazole + azoxystrobin, pydiflumetofen + difenoconazole, and difenoconazole were selected for compatibility tests. Each fungicide was tested at their recommended doses using a dilution method. Warm PDA around 60–65°C was mixed with each fungicide and poured into a sterile Petri dish to obtain the final concentrations of its recommended dose. Trichoderma plates were cultured on PDA for 7 d at 28 ± 2°C and a mycelial plug of 0.5 cm in diameter from each fungus was placed in the centre of prepared PDA plates and incubated for 3 days at 28 ± 2°C. The growth inhibition was calculated in comparison with the control.

The experiment was arranged at Zhongkai university of agriculture and engineering, South China. The Cavendish banana was used to demonstrate the efficacy of T. koningiopsis Tk905 in controlling banana Fusarium disease caused by FocTR4. The experiments were conducted on May 2020 and 2022. Five leaf stage Cavendish banana, aged about 2 months, were transplanted and planted in 9 cm-diameter pots, with nutrient soil (vermiculite: soil = 1:4). The experiment was conducted using a completely randomized design (CRD) with 10 plant seedling replicates. The entire experiment was conducted in triplicate. The treatments included three different types of inoculation: (1) banana plants inoculated only with FocTR4 spores (T-F+); (2) banana plants co-inoculated mycelium and spores of Tk905 + T + F and FocTR4 spores by root irrigation, mix with soil and spray (T+F+); (3) FocTR4 inoculated banana plants treated with pydiflumetofen + difenoconazole (PD + F+); (4) co-inoculated banana plants with Tk905 and FocTR4 (Mix soil), treated with pydiflumetofen + difenoconazole (T + PD + F+); (5) non-inoculated plants (T-F-). Tk905 was added 24 h prior to FocTR4 inoculation. A total of 8 mL Tk905 was put around the roots using by sterile pipette for the root irrigation experiment. After 10 days, a total of 25 mL FocTR4 spore suspension (3 × 10⁷/mL) was incubated around the roots. The mix with soil method was using the Tk905 cultured by solid-state fermentation. 3 g Tk905 fermentation was mixed with soil. Using the manual mini spray to spray the front and back of banana leaves was used for the spray method. The incidence rate and disease index were recorded 45 d after inoculation using the following formulas. Disease index = Σ ( number of leaves in each disease grade × grade value ) / ( total number of assessed leaves × highest grade value ) . Disease inhibition rate ( % ) = ( disease index of control − disease index after treatment ) / disease index of control × 100.

All data were statistically evaluated using ANOVA and means were separated by the Duncan’s multiple range test (p < 0.05 probability) using the Data Processing System (DPS, Tang and Zhang, 2013). Data were presented as mean ± standard error (SE). Graphs were constructed using Microsoft Excel.

The endophyte Trichoderma Tk905 strain isolated from the roots of the plant exhibited typical morphological features of T. koningiopsis based on the monograph of Samuels et al. (2006). Conidia were distributed on the aerial hyphae and eventually acquired a green later color. Conidial cells were oval to long oval, had smooth walls and a size of 2.7–5.3 μm × 2.0–3.5 μm (3.6 ± 0.6 × 2.7 ± 0.4 μm). Conidiophore upright, about 2 μm wide, with an obvious main axis. Conidiophores produced bottled stems or secondary branches directly. The angle between the branches and the main axis was slightly less than 90°. The tops of the fertile branches produced 2–5 bottle stems arranged in a vortex. Chlamydospores were not observed. The colony was initially white and after 2–3 days on PDA medium it started to produce spores and became lawn-like from the middle, with no obvious odor (Figures 1A–C).

The ITS sequence length of strain Tk905 was 604 bp (GenBank accession number OR083438). Tk905 sequence was up to 100% identical to known T. koningiopsis sequences belonging to strains CBS119075, MT111912, MN602617, MF616361, MH512944, KY111266, KC884758, EU718083, and DQ379015. The Tk905 Tef1-α sequence was 354 bp long (GenBank accession number: OR096250).

Phylogenetic analysis based on ITS and TEF-1α genes revealed that the Tk905 strain shared the highest homology in the phylogeny and clustered together with T. koningiopsis strains (Figure 1D). Based on the morphological and molecular data, the Tk905 strain was identified as T. koningiopsis and named T. koningiopsis Tk905.

In this study, Tk905 was cultured with A. alternata, C. ilicicola, Co. gloeosporioides, F. oxysporum, FocTR4 14,013, H. maydis, L. pseudotheobromae, and S. pogostemonis for 7 days and was able to inhibit them in varying degrees. The Tk905 mycelial growth inhibition rates ranged between 26.52–75.34% (Figure 2B). The Tk905 mycelium covered the pathogen mycelium and some Tk905 spores were observed on or around the pathogen hyphae (Figure 2A).

Furthermore, growth-promoting effects of T. koningiopsis Tk905 were evaluated based on fresh root height, plant height, and fresh weight measurements of the banana plants. The average root fresh weight, shoot fresh weight and plant height of the T. koningiopsis Tk905-inoculated plants were 3.96 ± 0.29 g, 12.51 ± 0.47 g, and 13.01 ± 0.27 cm, which were significantly higher than those of untreated plants with 2.42 ± 0.33 g, 9.96 ± 0.64 g, and 11.55 ± 0.21 cm, respectively (p < 0.05). Overall, the Tk905 strain significantly affected root fresh weight (38.89%), plant height (11.23%), and shoot fresh weight (20.38%) in bananas (Figure 3).

Upon inoculation with Tk905, notable changes were observed in the activity levels of the antioxidant enzymes CAT, PAL, PPO, and SOD in both leaves and roots of the banana plants (Figure 4). Recognized for its crucial role in scavenging free radicals in plants, the activity levels of POD in the roots were consistently elevated in treated plants across all time points. Notably, a peak was observed at 72 h post-treatment. By 24 h, POD levels in treated leaves and roots were 1.63 times and 2.17 times higher than the controls, respectively. For treated plants, there was a heightened PPO activity in both leaves and roots at all monitored time points. A peak activity, 3.6 times (leaves) and 1.58 times (roots) higher than the control, was recorded 12 h post-treatment. Maximum CAT activity was reached at 72 h in leaves and 12 h in roots. The root’s CAT activity under Tk905 treatment showed a remarkable increase of 21.27% at 48 h. Playing a pivotal role in plant defense under adverse conditions, PAL activity showed an upward trend in both leaves and roots post-Tk905 treatment. The leaves registered a significant 28.7% boost in PAL activity by 72 h (Figure 4).

All the seven tested agents had inhibitory effects on FocTR4-14013, and the inhibition rate ranged between 61.61–81.04% and 57.14–79.15% 3 and 7 d after fungicide treatment, respectively. On the third day of treatment of FocTR4-14013 with seven fungicides, the highest inhibition rate was found in the water dispersible difenoconazole, which reached 81.01%, followed by the pydiflumetofen + difenoconazole suspension agent which exhibited a 79.15% inhibition rate (Table 1).

The two most effective fungicides (difenoconazole and pydiflumetofen + difenoconazole), were selected for compatibility tests with Tk905. Although the above fungicides two had some limiting effects on TK905 growth at a concentration of 50 μg/mL, it could still grow, and the inhibition rate fell to approximately 20% after 8 d. Additionally, Tk905 can grow well in concentrations of 10 μg/mL and 20 μg/mL. Overall, results suggested that T. koningiopsis Tk905 was compatible with difenoconazole and pydiflumetofen + difenoconazole since it could grow normally on plates containing these fungicides at the recommended doses (Table 2).

To determine whether T. koningiopsis Tk905 could reduce the symptoms of FocTR4 infection in bananas, 5-leaf stage banana plant roots were treated with Tk905 using three different treatment methods: root irrigation, incorporation in soil, and spraying leaves using sterile fermentation broth as a dispersion medium and their susceptibility to FocTR4 (+T+F) leaf infection was subsequently assessed. MTk905-untreated plants (−T+F) and banana plants that were not treated with neither the pathogen nor the control agent (−T−F) were also included. Treatment of banana plants with Tk905 significantly reduced the disease index 45 days after inoculation (p < 0.05; Table 3).

Among different Tk905 application methods, the banana wilt disease control effect was found to be the best in the Tk905 irrigation (8 mL) treatment group reaching 52.41%, followed by the Tk905 incorporated with soil treatment group (49.6%). Furthermore, the combination of Tk905 with the pydiflumetofen + difenoconazole fungicide resulted in a synergistic effect: 71.40% control achieved by the combination while 52.92% by Tk905 and 57.10% by pydiflumetofen + difenoconazole alone (Table 3).