In this study, an F. verticillioides strain F01.12 (accession number MZ559332) isolated by Tran et al. (2021a) was used. In addition, five other Fusarium strains comprising F. proliferatum and F. mangiferae were isolated from maize grains in the central highlands of Vietnam (Tran et al., 2021b), F. oxysporum, and F. solani isolated from onion in Dalat, Vietnam (Le et al., 2020) and F. graminearum PH1 isolated from wheat (Gale et al., 2005) were also used. Each fungus was cultured on a PDA plate (Potato Dextrose Agar, 40 g L^–1) (Sigma Aldrich, Overijse, Belgium) for 7 days at 25°C before performing assays. To prepare a spore suspension of F. verticillioides, we placed a 7-day-old PDA plate of this fungus into a cabinet equipped with near-UV lights (12 h light/12 h darkness) for 7 days. The spore suspension was then collected and diluted with sterile water until a final concentration of 10⁷ conidia mL^–1.

We used four rhizobacterial strains ST02, ST03, ST07, and ST08 isolated from soil in Ben Tre, Vietnam. For the identification of these strains, we amplified the recombinase A (recA) gene by using a primer pair of recAF and recAR as described by Guo et al., 2008 (Supplementary Table 1). The sequences were deposited to the NCBI gene bank, representing MZ614615 for strain ST02, MZ614616 for strain ST03, MZ614619 for strain ST07, and MZ614615 for strain ST08. For inoculation, each strain was grown on a TSA plate (Tryptic Soy Agar, 30 g L^–1) (Sigma Aldrich, Overijse, Belgium) for 5 days. Three rhizobacterial colonies were then transferred into a 50-mL Falcon tube containing 20 mL TSB (tryptic soy broth, 30 g L^–1) (Sigma Aldrich, Overijse, Belgium) and incubated for 7 days at 200 g, and 28°C. The Falcon tubes were then centrifuged at 10000 g for 10 min and the cell suspension was obtained by discarding the supernatant through a four-fold Mira cloth followed by re-suspending the cell pellets in 5 mL sterile water. A ten-fold serial dilution was made to determine the colony-forming unit (CFU) of each cell suspension (5 × 10⁶ CFU mL^–1). The cell-free supernatant was concentrated 10 times using a nitrogen flow at 1 atm in a water bath at 40°C for 2 h.

In agar plates, we used two different models: (1) a diffusion assay and (2) a volatile assay. For the diffusion assay, we used a four-hole TSA plate (Φ 90 mm) (Supplementary Figure 1A). In each plate, a 7-day-old agar plug of each pathogen (Φ 5 mm) was placed in the center, and 50 μL of each rhizobacterial cell suspension or supernatant was filled in each hole. The plate was incubated at 28°C and the radius of the fungal colony was measured. In addition, to evaluate the effects of the volatile compounds produced by rhizobacterial strains we performed the volatile assay in which a TSA plate was divided into two parts by making an agar-free channel in the middle (Φ 5 mm) (Supplementary Figure 1B). Water and blank TSB were used as control.

In this assay, we used a six-well plate containing 4.5 mL TSB per well (Supplementary Figure 1C). A ST03 cell suspension (5 × 10⁷ CFU mL^–1) or a 10-time-concentrated supernatant was injected to each well at different volumes (μL) of 0 (0%), 31.25 (0.6%), 62.5 (1.3%), 125 (2.5%), 250 (5%), and 500 μL (10%). TSB was subsequently added to each well till 5 mL. A 7-day-old agar plug of Fusarium verticillioides (Φ 5 mm) was then transferred to each well. The plate was inoculated at 28°C for 7 days. To analyze the fungal growth on the surface, an image was taken using a custom-build multispectral phenotyping platform as described by Tan et al. (2020). This platform is equipped with a 6 Mp-16 bit camera mounted on a Cartesian coordinate robot and allows 6-μm-high-resolution multispectral imaging. The multispectral camera can visualize the surface growth of the fungus. Finally, biomass from each well was collected followed by freeze-dried and weighted.

This bioassay was employed using baby maize cobs (70 mm) (Baby corn, Excel Fruits, Thailand). In each treatment, five maize cobs were disinfected with 1% NaOCl solution, washed twice with sterile water, and then placed in a one-well plate equipped with a moist Whatman paper (Supplementary Figure 1D). Subsequently, we made a hole in the middle of each cob (depth ∼ 10 mm) using a sterile 10 μL pipette tip. A mixture consisting of 5 μL spore suspension of Fusarium verticillioides (10⁷ conidia mL^–1) and 5 μL cell suspension (5 × 10⁶ CFU mL^–1) [FV + ST03 treatment)/or cell-free supernatant of rhizobacterial strain ST03 (FV + Sup) treatment] was loaded to each hole. Each maize ear was inoculated at 28°C for 7 days. Five infected maize cobs were treated with sterile water as a positive control (FV + W) and five F. verticillioides non-infected cobs were treated with sterile water as a negative control (W + W). The cobs were then divided into two equal parts using a scalpel. To determine the damage level caused by the pathogen, we analyzed images taken by the multispectral camera using the RGB module. Finally, each maize cob was thoroughly ground using liquid nitrogen for quantification of FB₁, FB₂, and FB₃ by an LC-MS/MS method (De Boevre et al., 2012).

In this assay, we used two hybrid maize lines Bt/GT NK7328 (supplied by Syngenta company, Vietnam) and CP888 (supplied by C. P. company, Vietnam) as they are predominantly planted in the central highlands of Vietnam (Tran et al., 2021c). Maize seeds were disinfected with 1% NaOCl solution and germinated in a rectangular plastic pot (20 × 16 × 6 cm) with vermiculite substrate (Vermex, Soprema, Belgium) for 5 days. For the biocontrol treatment (FV + ST03 treatment), five germinated seeds were infected with F. verticillioides by soaking in a fungal suspension (10⁷ conidia mL^–1) for 1 h, and then sowed in five glass tubes (a seed per a tube) containing soil inoculated with the ST03 cell suspension. To inoculate the soil with Streptomyces, 3 g polymer gel (DCM Aquaperla, Grobbendonk, Belgium) absorbed in 300 mL tap water was mixed with 2 kg non-sterile fine sand (Voss chemie-Benelux, Lier, Belgium) (De Zutter et al., 2021). This mixture was then inoculated with the ST03 cell suspension in a concentration of 10⁶ CFU g^–1 soil 3 days before planting. The positive control comprised five F. verticillioides infected seeds in soil non-inoculated with the ST03 cell suspension (FV + W treatment). For the negative control, five mock seeds treated with sterile water were planted in non-inoculated soil with the strain ST03 (W + W treatment), and five mock seeds were planted in soil pre-inoculated with ST03 (W + ST03 treatment). Every 2 days plants were watered with 10 mL tap water. To assess the direct antagonism of strain ST03, the soil was collected after 2 weeks of planting and the DNA levels of F. verticillioides were qualified as described below.

To measure the impact of ST03 on the plant growth when co-applied with F. verticillioides in maize seedlings, the length and fresh biomass of leaves and roots were measured at 16 days after inoculation (dai). Chlorophyll Index (ChlIdx) was also assessed using the multispectral camera as described by Tan et al. (2021). Meanwhile, to uncover the impact of the Streptomyces strain ST03 on the expression of genes related to plant defense response, plants of each treatment were harvested at four different time points at dai 1, 2, 4, and 8 (five plants per each time-point) (Supplementary Figure 1E). Roots and leaves from each plant were extracted for RNA as described below to quantify the gene expression using RT-qPCR.

Total RNA from roots and leaves was extracted using TRizol reagent (Sigma Aldrich, Belgium) according to the manufacturer’s instructions. RNA concentration was determined using a Quantus fluorometer (Promega, Netherlands). cDNA of each RNA template was synthesized using the iScript^TM kit (Bio-Rad, Belgium) and diluted five times with nuclease-free water. The quantitative reverse transcription PCR (RT-qPCR) assay was performed using a CFX96 Tough Real-time PCR Detection System (Bio-Rad, Belgium). Each PCR reaction contained 6.3 μL Gotagq^®PCR master mix (Promega, Netherlands), 2 μL cDNA, 0.6 μL each primer (5 μM), and 0.2 μL CXR dye (Promega, Netherlands), and 2.3 μL nuclease-free water. The thermal program was set up as follows: 95°C for 3 min; 39 cycles of 95°C for 10 s, and 60°C for 30 s, followed by a melting curve acquisition from 65 to 95°C with the rate of 0.5°C s^–1. The primers used for all target genes are shown in Supplementary Table 1. Elongation factor 1α (EF-1α) and β tubulin (β-TUB) primers were used as housekeeping genes. Gene expression analysis was done using qBase⁺ software (Biogazelle, Zwijnaarde, Belgium). Fold change was computed by diving the CNRQ values (calibrated normalized relative quantities) of the treated samples by values of the control samples. In each gene, four biological replicates and two technical replicates were done.

Quantification of fumonisin B₁, B₂, and B₃ in each maize ear was performed using a Waters Acquity HPLC-Quattro Premier XE mass spectrometer (Waters, Milford, MA, United States) in positive electrospray ionization (ESI⁺) mode. Chromatographic separation was performed using a Symmetry C18 (150 mm × 2.1 mm, i.d. 5 μm) column with a guard column (10 mm × 2.1 mm i.d.) of the same material (Waters, Zellik, Belgium). A clean-up procedure was carried out prior to injecting samples into the device. Briefly, 100 μL de-epoxy-deoxynivalenol (DOM) (50 ng μL^–1) were added to a Falcon tube containing 5 g of each ground sample as an internal standard. Each sample was extracted with 10 mL ethyl acetate/folic acid (99/1, v/v) by agitating for 15 h on an overhead shaker. The Falcon tubes were centrifuged at 4000 g for 15 min. Each upper layer was transferred into a new tube and evaporated using a nitrogen flow at 40°C, 1 atm. The pellet was re-dissolved in 200 μL injection solvent {consisting of mobile phase A [water/methanol/acetic acid (94/5/1, v/v/v) + 5 mM ammonium acetate] and mobile phase B [methanol/water/acetic acid (97/2/1, v/v/v) + 5 mM ammonium acetate] with a ratio of A/B (60/40, v/v)}; and the mixture was defatted with 200 μL hexane and subsequently filtrated using a centrifuge filter at 10000 g for 5 min. Finally, each filtrate was conveyed into an HPLC vial. The liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) method was described by De Boevre et al. (2012). Five biological replicates were done per treatment.

A quantitative PCR (qPCR) assay was conducted in a CFX96 Touch Real-Time PCR Detection System (BIO-RAD, Temse, Belgium) as described by Tran et al. (2021a). Briefly, total genomic DNA extraction was performed from 100 mg for each soil sample, and 50 mg for pure F. verticillioides mycelia using the E.Z.N.A^® Soil DNA Kit (VWR International, Leuven, Belgium). DNA concentration was quantified with a Quantus fluorometer (Promega, Leiden, Netherlands). Each reaction mixture contained 6.3 μL Gotag^® qPCR Master Mix (Promega, Leiden, Netherlands), 2 μL DNA template, 0.6 μL each primer FVer (5 μM), 0.2 μL CXR reference dye (Promega, Leiden, Netherlands), and nuclease-free water up to a total volume of 12 μL. The PCR thermal cycling program was set-up as follows: 95°C for 3 min; 39 cycles of 95°C for 10 s, and 60°C for 30s, followed by a melting curve acquisition from 65 to 95°C. A standard curve was generated based on C_t (threshold cycle) values by using tenfold serial dilutions of the pure F. verticillioides DNA (ranging from 1 ng to 1 × 10^–5 ng/μL). The amount of fungal DNA in samples was calculated from C_t values using the standard curve. Each sample calculation was performed twice. Five biological replicates were undertaken per treatment.

All heat maps and boxplots were generated using the R software v.4.0.2 with the packages ggplot2 and gplots¹. The normal distribution of data was tested using the Shapiro–Wilk test. The one-way ANOVA test followed by a post hoc Tukey test was used in a case of the normal distribution, otherwise, a non-parametric Kruskal–Wallis test and a post hoc Dunn’s test were applied. As for statistical analyses in gene expression, fold change (FC) was used in the case of up-regulation, while for down-regulation log [FC] base 2 (Log₂FC) was used. All analyses were tested at a significance level of α = 0.05.

In this study, all the rhizobacterial strains were identified as Streptomyces species. The in vitro bioassays showed an effective and stronger direct antagonism of the strain ST03 against six Fusarium pathogens compared to the other rhizobacterial strains (Figure 1A). It is clear that in the in vitro diffusion assay, an early and significantly inhibitory indication was observed against F. verticillioides, F. proliferatum, F. oxysporum, and F. graminearum when co-applied with the strain ST03 at 2 dai and at 3 dai for F. mangiferae and F. solani in comparison to the mock control. Surprisingly, F. verticillioides, F. proliferatum, and F. oxysporum were thoroughly inhibited by the strain ST03 from 2 dai onward since their colony radii remained unchanged between 2 and 7 dai. It was coherent when its inhibition efficacy against these three pathogens was highest and significant at 7 dai compared to the other bacterial strains (Figure 1B and Supplementary Figure 2). The results also indicated absolute but a later inhibition of F. mangiferae, F. solani, and F. graminearum when co-inoculated with the strain ST03 at 5 dai (Figure 1A), which explained lower inhibition efficacy at 7 dai, of 65 ± 1%, 56 ± 2%, and 51 ± 2%, respectively (Figure 1B and Supplementary Figure 2). It was significantly higher than the inhibition efficacy of the other strains ST02, ST07, and ST08. An obvious example is that at 7 dai the inhibition efficacy against F. graminearum by the strain ST03 was 51 ± 2%, while this was not significantly effective for the other ones, for example, ST02 (3 ± 2%, p < 0.001), ST07 (6 ± 8%, p < 0.001), and ST08 (4 ± 2%, p < 0.001) (Supplementary Figure 2). Remarkably, the antagonistic efficacy of the strain ST03 reached above 50% at 5 dai for all the pathogens except for F. solani (Supplementary Figure 2). We, therefore, used this strain ST03 for in-depth studies.

To assess the impact of volatile compounds produced by strain ST03, we used an in vitro volatile bioassay. Even though this assay did not indicate the role of volatile metabolites for the antagonism against the pathogens, it strengthened the diffusion-based antagonism since the inhibition zones were more transparent for each pathogen compared to the mock control (Figure 1C). At 11 dai, F. verticillioides, F. proliferatum, and F. oxysporum did not grow onto the other side inoculated with strain ST03, whereas for the mock control plates the fungi colonized the other side from the channel by approx. 25 mm. As similar to the diffusion assay, the inhibition efficacy was less for F. mangiferae, F. solani, and especially for F. graminearum (Figure 1C).

The high antagonistic efficacy of strain ST03 was proven in the other in vitro assay using a liquid medium. It was confirmed by a significant decrease in biomass production by F. verticillioides when co-inoculated with the cell suspension of strain ST03, ranging from 2.5 to 10% (Figure 2A). The fungal growth was remarkably suppressed when co-applied with 2.5% of strain ST03 equivalently with 1.25 × 10⁶ CFU mL^–1 because there was a significant reduction of dried biomass by 59% compared to the mock control (0% treatment), specifically 13 ± 1 and 32 ± 9 mg, respectively (p = 0.021). This trend also occurred in the treatments co-inoculated with 5% (12 ± 1 mg, p = 0.017) and 10% (8 ± 2 mg, p = 0.005) of the cell suspension, but the change was not different from the treatment 2.5% (Figure 2A).

We hypothesized whether extracellular compounds secreted by strain ST03 in the supernatant are involved in the interaction with the pathogens. To explore this, we concentrated the supernatant 10 times after separating from the cell and applied both in vitro bioassays in the same way with the cell suspension. As indicated in the plate diffusion assay (Figure 1A), the supernatant of strain ST03 had a significantly effective antagonism against six Fusarium species. It was clear that the growth of all fungi were strongly inhibited when co-applied with the supernatant at 3 dai by their smaller radii compared to the mock controls. For example, the colony radius for F. verticillioides was 10 ± 2 cm vs. 21 ± 1 cm (p < 0.001), F. proliferatum (10 ± 0.5 cm vs. 18 ± 0.3 cm, p < 0.001), and F. graminearum (13 ± 1 cm vs. 30 ± 1 cm, p < 0.001). Surprisingly, an absolute inhibition was observed at 3 dai for all the pathogens with the exclusion of F. solani when the fungi did not grow further after 3 days of being inoculated with the supernatant (Figure 1A). This phenomenon resulted in a high inhibition efficacy against these pathogens of over 60% at 5 dai and over 70% at 7 dai (Supplementary Figure 2).

In addition, the antifungal activity of this supernatant was demonstrated in the liquid assay (Figure 2). The fungus F. verticillioides was significantly inhibited when co-applied with 0.6% of the supernatant at 7 dai by a significant decrease in dried biomass compared to the mock control (0%), given 17 ± 5 mg vs. 29 ± 6 mg, respectively (p = 0.016) (Figure 2A). Although no significant change in the biomass production of the fungus in a range of between 0.6 and 2.5% was observed, more suppression was observed at 5% (8 ± 3 mg dried biomass, p < 0.010), and 10% (3 ± 2 mg dried biomass, p < 0.001) than 2.5% (14 ± 2 mg) (Figure 2A). Moreover, it was clear that the growth surface of the fungus was remarkably smaller in the treatments 2.5, 5, and 10% compared to the mock control (Figure 2B). Moreover, an inhibition efficacy calculated at a concentration of 10% was 90 ± 8%.

In conclusion, the in vitro antagonistic bioassays show that of the four Streptomyces strains, strain ST03 had the highest inhibition efficacy against all the surveyed Fusarium pathogens. The data also indicate the antifungal potential of the extracellular compounds in the supernatant produced by this strain.

To confirm the antagonistic efficacy of strain ST03 from the in vitro parts, we performed two in vivo bioassays on maize cobs and maize seedlings infected with F. verticillioides. We focused on F. verticillioides since it is the main causal agent of FER and predominantly occurs in maize fields worldwide.

First, we assessed the impact of strain ST03 when co-applied with F. verticillioides on maize cobs (Figure 3). Data illustrated that the infection of the fungus to maize cobs was significantly counteracted when co-inoculated with the cell suspension or supernatant of strain ST03 (Figure 3). By using the multispectral camera, a high infection level of F. verticillioides on maize cobs (red area) (FV + W) was observed with a larger lesion area when compared to the negative control (W + W) (p = 0.001) (Figures 3A,B). However, it significantly reduced when co-applied with the cell suspension (FV + ST03) or supernatant (FV + Sup) of strain ST03 by smaller surface areas (p = 0.008), respectively (Figure 3B).

For the in vivo bioassay on maize seedlings, we checked the direct antagonism of the biocontrol strain ST03 by quantification of the F. verticillioides DNA in soil surrounding roots upon 14 days of infection. We planted F. verticillioides-infected CP888 seeds into two types of soil with and without pre-inoculation with the biocontrol strain ST03 (Figure 4A). qPCR data indicate that a higher level of F. verticillioides DNA was present in soil without pre-inoculation with strain ST03 (FV + W), compared to soil that was pre-inoculated (p = 0.009) (Figure 4B). This was consistent with the phenotypic characteristics of roots and lower levels of infection were observed in plants treated with the biocontrol strain compared to the non-treated plants (Figure 4A). All evidence shows that strain ST03 could suppress F. verticillioides through a direct antagonism upon infection on maize seedlings.

To sum up, both in vivo bioassays on maize cobs and plants demonstrate the direct antagonism of the biocontrol strain ST03 against F. verticillioides. In addition to cell suspension, the extracellular compounds in the supernatant secreted by this strain also show a strong inhibition against F. verticillioides.

The in vivo antagonistic bioassay on maize cobs indicated the inhibition efficacy of both cell suspension and supernatant of the strain ST03 by the reduction of the necrosis area (Figure 3). To evaluate the impact on fumonisin production by F. verticillioides we measured FBs in these cobs. Data show that levels of FB₁, FB₂, and FB₃ in maize cobs treated with the cell suspension (FV + ST03) or the supernatant (FV + Sup) was significantly lower than in the non-treated maize cobs (FV + W) (Figure 5). A significantly lower FB₁ median level was found in maize cobs treated with the cell suspension compared to the non-treated maize cobs, given 9 ± 5 μg kg^–1 and 80 ± 10 μg kg^–1, respectively (p = 0.012). A similar result was observed for FB₂, amounting 0 ± 4 μg kg^–1 (in FV + ST03 treatment) and 92 ± 25 μg kg^–1 (in FV + W treatment) (p = 0.009), and for FB₃ (0 ± 2 μg kg^–1 vs. 22 ± 3 μg kg^–1, p = 0.013). Similar results were also observed for the maize cobs treated with the supernatant (FV + Sup) except for FB₂ (Figure 5). Collectively, the biocontrol strain was capable of suppressing the production of fumonisins by F. verticillioides.

After uncovering the direct antagonism of the biocontrol strain on maize seedlings by the reduction of the DNA level of F. verticillioides in soil (Figure 4), we explored whether the growth of plants was impacted. To examine this, we used two hybrid maize lines CP888 and Bt/GT NK7328 predominantly planted in the central highlands of Vietnam (Tran et al., 2021c). For line CP888, the growth of plants infected with the fungus (FV + W) was negatively affected with significantly lower fresh leaf and root biomass weight when compared to the control non-infected plants (Figure 6A). Moreover, a reduction of chlorophyll index was observed in the infected plants versus the control plants (Figure 6B). For line Bt/GT NK7328, although the plants were less infected compared to line CP888, their root length was also significantly shorter than the control NK7328 plants, of 10.3 ± 0.3 and 11.8 ± 0.3 cm, respectively (p = 0.029) (Figure 6A).

However, when co-applied with the biocontrol strain ST03 (FV + ST03), the growth of plants was significantly stimulated compared to the infected and the control plants of both maize lines. For example, for line CP888, the leaf length of plants co-applied with strain ST03 was 22.5 ± 1.4 cm (FV + ST03) versus 13.0 ± 1.0 cm of the infected plants (FV + W) (p = 0.029). Likewise, for Bt/GT NK7328 plants the length was 39.4 ± 1.0 cm (FV + ST03) vs. 30.2 ± 3.1 cm (FV + W) (p = 0.029) (Figure 6A). In both cultivars, the growth of the plants when co-applied with the biocontrol strain was also better than the growth of the control plants (W + W). It was evidenced by higher fresh biomass and leaf length (Figure 6A). Chlorophyll index was higher in the leaves of CP888 plants when co-applied with strain ST03, given 2.3 ± 0.1 (FV + ST03) vs. 1.4 ± 0.1 (FV + W) (p = 8e-4) or vs. 1.6 ± 0.1 (W + W) (p = 0.008) (Figure 6B).

We hypothesized whether this biocontrol strain ST03 acts as a plant growth-promotion agent on top of its direct antagonism. To investigate this, we examined growth indicators of plants that were applied with strain ST03 alone (W + ST03). Data revealed that the plant growth of both lines was significantly promoted with the presence of ST03 strain compared to the control plants (W + W). For example, in Bt/GT NK7328 line, leaf length was longer at 40.1 ± 0.5 (W + ST03) vs. 28.4 ± 3.1 cm (W + W) (p = 0.029), and a higher leaf biomass of 2.1 ± 0.1 vs. 1.3 ± 0.1 g (p = 0.029) (Figure 6A). Similar results were observed in plants of line CP888 (Figure 6A). The chlorophyll index of plants applied with strain ST03 was significantly higher than that of the control plants, given 2.2 ± 0.1 and 1.6 ± 0.1, respectively (p = 0.008) (Figure 6B). Nevertheless, the primary root length of plants applied with the biocontrol strain was shorter compared to the control. It was true in both lines CP888 [of 9.5 ± 0.2 (W + ST03) vs. 11 ± 0.6 (W + W), (p = 0.145) and Bt/GT NK7328 (of 10.0 ± 0.6 vs. 11.8 ± 0.3, (p = 0.028)]. Higher fresh root biomass was observed in plants applied with strain ST03 compared to the control plants (Figure 6A) since root hair formation was induced as shown in Figure 4A.

From the aforementioned data, we can conclude that the Streptomyces strain ST03 not only hampered the growth of F. verticillioides but also stimulated the growth of plants.

Concerning the action of this biocontrol strain on modulating plant growth promoting hormones, the involvement of genes associated with the metabolic pathways of auxin/indole-3-acetic acid (Auxin/IAA), abscisic acid (ABA), and gibberellic acid (GA) were examined. In the literature, the timely activation of key genes in the biosynthesis pathways of salicylic acid (SA), jasmonic acid (JA), 1,4-benzoxazine-3-ones (BXs), and pathogenesis-related proteins (PRs) plays a dominant role in maize resistance to fungal plant pathogens (Bacon et al., 2007; Ding et al., 2015; Wang et al., 2016). Thus, we further uncovered modes of action of the biocontrol strain ST03 on these secondary metabolism-related genes.