Previously identified fungus A. niger (obtained from College of Chemistry and Life Sciences, Chengdu Normal University, Chengdu, China) was used for the extraction of antibacterial compound. Potato dextrose agar and potato dextrose broth were used for culturing and for the seed broth preparation of A. niger. Solid-state fermentation (SFM) was used for A. niger fermentation. Beef extract peptone broth (BEPB) and agar medium were used for culturing R. solanacearum and for antimicrobial tests. The composition of each medium is presented in Supplementary Table S1 . Bacterial wilt pathogen R. solanacearum (biovar 3, race 1) was cultured for its logarithmic phase on beef extract peptone solid medium at 30°C and 150 rpm/min for 16 h (Liu et al., 2016). The activated strain was grown to its logarithmic phase in 35 ml of BEPB under shaking incubation at 150 rpm/min and 30°C. The culture of the bacteria was centrifuged at 6,000 rpm, and by using sterilized distilled water the concentration of 10⁸ CFU/ml was maintained.

Shallow plate fermentation method was used for fungal fermentation. SFM was inoculated with 15% (v/m) spore suspension of A. niger with 1 × 10⁶ spores/ml concentration and incubated for 1 week at 28°C. After the production of spores, the solid matrix was dried for 36 h and ground into particles. The spore powder of A. niger was collected, and the spore concentration was measured as 2 × 10⁸ spores/g. Crude ethanolic extract of spore powder was prepared by extracting spore powder four times for 4 h at 80°C using 10 L ethanol each time. In 300 ml of distilled water, 120 g of the extract was dissolved and then successively extracted four times with 300 ml of ethyl acetate (EA), petroleum ether (PE), and water to produce EA, PE, and water fractions. Separation of crude fractions of PE was done under the regulation of antimicrobial activity. Using silica gel column chromatography, the PE extract was fractionated, and the gradient was eluted. Thin-layer chromatography (TLC) analysis was used to evaluate the polarity composition of the fractions (200 ml), and those that have a similar composition were then concentrated and combined. The TLC analysis of the PE extract showed that it comprises different metabolites having various levels of bioactivity. Thus, additional fractionation was required to obtain more metabolites. Lastly, this separation procedure produced eight fractions (F1–F8) ( Figure 1 ). The antimicrobial potential-guided separation of the most active fraction F8 was performed using reversed-phase silica gel column chromatography (CC), eluted with water–acetone (1:0, 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8, 1:9, and 0:1) to yield five sub-fractions (F8-1, F8-2, F8-3, F8-4, and F8-5). F8-1 (4.634 g) was again subjected to CC using silica gel (300–400 mesh) and eluted through petroleum ether–acetone (10:4). The further purification of sub-fraction was performed through silica gel CC with petroleum ether–acetone (5:2, v/v) as eluent to yield a white crystal compound (360 mg). The compound was subjected to liquid chromatography–mass spectrometry (MS), electron ionization MS, and NMR (¹H-NMR and ¹³C-NMR) spectroscopic analysis, and by matching the obtained data with already published data, the structure of the compound was determined.

The crude extract and the fractions were tested for antibacterial activity against BW wilt pathogen R. solanacearum. At first, 100 mg/ml concentration of crude extract and active fractions were prepared separately, and then 10 μl of this concentration was poured to 1 ml of BEPB. Furthermore, this was inoculated with 100 μl R. solanacearum suspension [10⁸ colony-forming units (cfu)/ml] and incubated under shaking incubation at 28°C and 160 rpm to logarithmic phase. Dimethyl sulfoxide was kept as solvent control, and streptomycin was used as the positive control. Using a UV–vis spectrophotometer (Shenyang Ebetter Optics Co., Ltd.), the bacterial cell density was monitored at OD_λmax (420 nm). The bacterial inhibition rate I_R was calculated by using the following formula: I_R = (OD₀ – OD₁)/OD₀ × 100. OD₀ and OD₁ represent the OD_λmax of treatment and blank control, respectively.

Destruction in bacterial cell morphology caused by the most active antibacterial compound (FA) was investigated through scanning electron microscopy. Bacterial cells under the treatment of FA and control (without treatment) were fixed using 2.5% glutaraldehyde and phosphate buffer for 1.5 h at 45°C, followed by washing with phosphate buffer (0.1 M, pH 7.2) for 5 min and fixing in osmium tetroxide (OsO₄) for 1 h. Using phosphate buffer, the samples were washed, dehydrated in graded ethanol series (20%, 40%, 60%, 80%, and 90%) for 10 min each, and subjected to ethanol and CO₂ to reach the critical point. By using gold ions, the samples were coated and subjected to scanning electron microscopy (SEM) evaluation (Kamonwannasit et al., 2013).

Tomato plants (25 days old) were transplanted into sterilized culture medium pots (one plant per pot). Four treatments were maintained: T1, control plants without FA treatment and inoculation of pathogen (Ck); T2, plants inoculated with R. solanacearum suspension (10⁸ cfu/ml) (Rs); T3, plants treated with 9 ml FA suspension (300 μg/ml) (FA); and T4, plants inoculated with 15 ml of pathogen suspension (10⁸ cfu/ml) and treated with 9 ml FA suspension (300 μg/ml) (Rs+FA). According to the methods described by Kiefer et al. (2000), total RNA was isolated from the tomato plants. Briefly, 0.2 g of plant material was ground in liquid nitrogen to obtain tissue powder in which 1,000 μl TRIzol reagent was poured and mixed. The samples were incubated for 8 min on ice, followed by addition and mixing of 200 μl chloroform. The mixture was then centrifuged at 12,000 rpm for 12 min at 6°C. The obtained pellet was washed using ethanol (75%), mixed in 30 μl RNAse-free water, and stored at −80°C. The expression of defense-related genes (PAL, LOX, PR1, and PR2) was evaluated through real-time polymerase chain reaction (RT-PCR) using the isolated RNA from tomato plants under different treatments. The primers used for the analysis of defense-related genes are presented in Supplementary Table S2 . For reference, Ubi3 was used. The RT-PCR reaction was conducted in 25 μl reaction mixture on DNA Engine Opticon2 Continuous Fluorescence Detection System. The composition of the reaction mixture and the operating conditions are given in Supplementary Table S3 . RNAse-free water was used instead of cDNA for negative control. Following the extension stage, the fluorescence signal was evaluated immediately after 2-s incubation at 70°C, which prevents the possibility of primer dimer detection. The melting points of the PCR products were measured between 65°C and 95°C at the end of the cycles. Agar gel electrophoresis and melting curve analysis were used to confirm the specificity of the amplicons. The experiment was conducted in three replicates.

A pot experiment was conducted in a greenhouse for evaluating the potential of FA to manage the bacterial wilt of tomato. Plastic pots (20 cm in diameter) were filled with 1 kg soil (sand/clay/silt = 25%:25%:50%) each, and 20-day-old tomato plants (cultivar Rio Grande) were transplanted (one plant/pot). In each pot, 15 ml of pathogen suspension (10⁸ cfu) was added to initiate the bacterial wilt disease. After 3 days of inoculation with the pathogen, the rhizosphere of each plant was treated with 9 ml of sterilized water containing an increasing concentration of FA (control and 50, 100, 150, and 200 μg/ml). The treatments were applied through three holes surrounding the plant. Following horticultural recommendations, the plants were irrigated and fertilized for 50 days. Each treatment was applied in seven replicates using completely randomized design (CRD). The experiment was terminated after 50 days, and data were taken on the pathogen population in soil, disease severity, and plant growth parameters (plant length, root length, and plant biomass). Data on pathogen population and disease severity were converted to log₁₀ and area under the disease progress curve (AUDPC) value, respectively, according to a previously described method (Khan et al., 2020). The experiment was repeated once, and the data of the two experiments were tested for significant difference and, in case of no significant difference, pooled for analysis.

Field evaluation of FA compound for controlling BW disease in tomato was tested in an experimental field at the College of Chemistry and Life Sciences, Chengdu Normal University, Chengdu 611130, China, in March 2021 and repeated in March 2022. The field area was divided into three sets of blocks, and every set contained three blocks, each with a size of 6 m². Tomato seedlings (20 days old) were transplanted in all blocks by maintaining two rows of tomato plants per block and eight plants in each row. The base of each plant in the two sets were treated with 100 ml of pathogen suspension (10⁸ cfu/ml) mixed with 300 ml of water. The plants in one of these two sets were treated with 20 ml of FA suspension (400 μg/ml) at the rhizosphere (T1: Rs+FA), while the other set was kept as untreated inoculated control (T2: Rs). The plants in the third set of blocks were kept as untreated and un-inoculated control (T3: control). The experiment was terminated after 50 days, and data were taken on soil bacterial population, disease severity, and plant growth parameters (plant height, root length, and plant fresh biomass). Data on disease severity were converted to AUDPC value according to a previously described method. Soil bacterial population was calculated twice: once at the start after 24 h of pathogen inoculation (initial: Pi) and once at the end of the experiment (final: Pf). The difference between Pi and Pf was calculated and expressed as decrease in soil bacterial population. The experiment was repeated once, and the data of the two experiments were tested for significant difference and, in case of no significant difference pooled for analysis.

The experiments were conducted using CRD in the laboratory and greenhouse, while field tests were evaluated using randomized complete block design. The data were analyzed through Statistical Analysis System software version 8.0. Analysis was done using one-way analysis of variance. Tukey’s multiple-range test (P < 0.05) was used to evaluate the significant differences among treatment means.

Different solvent extracts of A. niger were tested for the inhibition of R. solanacearum growth. Among different extracts, the petroleum ether extract showed the highest antibacterial activity that was equal to the positive control streptomycin. The inhibition rate of petroleum ether extract and streptomycin was 89.4% and 91.3%, respectively, followed by ethyl acetate extract and aqueous extract that showed 39.5% and 23.6% inhibition rate, respectively ( Figure 2A ). Based on the maximum inhibition rate, petroleum ether extract was selected for further fractionation and analysis of antibacterial compound. The fractionation of petroleum ether extract resulted in eight fractions showing varying degrees of inhibition rate against R. solanacearum. Among these eight fractions, F4, F5, and F8 showed a clearly higher antibacterial potential compared with others, and among these three fractions, significantly highest inhibition rate of 90.2% was shown by F8, which was similar to that obtained by positive control which indicated the presence of antibacterial compounds in these three fractions and especially in F8 ( Figure 2B ). As the fractions F4 and F5 showed inhibition rates of less than 60% and F8 only exhibited more than 90% inhibition rate, F8 was further processed for separation and purification of the antibacterial compound. Finally, a compound of white crystal nature was obtained. Based on their spectroscopic data, it was identified as FA compound—5-(hydroxymethyl)-2-furoic acid—with molecular formula C₆H₆O₄ ( Supplementary Figures S1, S2 ).

Bacterial cell morphology under treatment of FA compound and without treatment was observed in SEM analysis. The micrographs clearly showed severe destructions in morphology of FA-treated bacterial cells. Compared with the control ( Figure 3A ), the FA-treated cells were swollen and the membranes were disrupted ( Figure 3B ). Leakage of cell content was obvious in the damaged cells, whereas in the untreated control the cells were in uniform morphology with a plain rod shape.

The expression of host defense-related genes (PAL, LOX, PR1, and PR2) in tomato plants was analyzed through RT-PCR after 30, 60, 90, and 120 h of pathogen inoculation under FA treatment ( Figure 4 ). The results showed that the FA-treated pathogen-inoculated plants showed an induced expression of resistance genes (PAL, LOX, PR1, and PR2) more than the basal level of expression in other groups where the plants were treated with pathogen alone, FA only, or un-inoculated and untreated control pants (Ck). The induction of PAL and PR1 genes was highest at 5.7- and 20-fold, respectively, at 120 h, while genes LOX and PR2 showed a maximum expression of 9.5- and 22-fold at 90 h, respectively, although plants with only pathogen inoculation or treated only with FA also showed an induced expression of these genes as compared with the control, but at a significantly lower rate and fold than the pathogen-inoculated plants that were treated with FA.

Consistent with the greenhouse results, the application of FA also reduced the bacterial wilt disease severity and soil population of R. solanacearum in field evaluation ( Figure 6 ). The plant growth was also enhanced by the application of FA ( Table 2 ). Compared with untreated inoculated control plants, the plants treated with FA exhibited a significantly lower AUDPC value (1,670) and the highest decrease (11.2 log₁₀ cfu/g) in soil population of R. solanacearum. The inoculated plants under the treatment of FA had improved plant length (58 ± 2.7 cm), plant biomass (28 ± 1.7 g), and root length (13 ± 1.2 cm) as compared with the untreated inoculated plants. The results of the FA treatment showed a similar plant growth as that recorded for untreated and un-inoculated healthy control plants.