Wild-type tomato (Solanum lycopersicum) cultivars “Castlemart” and “Moneymaker”, the JA biosynthetic mutant spr2 in the “Castlemart” cultivar (Li et al., 2003), and the SA-deficient NahG transgenic line in the “Moneymaker” cultivar were kindly provided by Prof. Zhao Jiuhai (Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences). The susceptible tomato cultivar “Zhongshu-4” was used in all nematode inoculation assays, unless indicated otherwise. Tomato seeds were surface sterilised in 1% NaClO for 5 min and then rinsed thoroughly with sterile water three times. The seeds were germinated in sterile vermiculite for 5–7 days at 26°C and then maintained in a growth chamber with a photoperiod of 16-h light (26°C) and 8-h dark (21°C). After 3 weeks of growth, tomato plants were used for nematode inoculation. All plants were watered daily and fertilised twice per week with Hoagland solution.

The population of Meloidogyne incognita used in this study was cultivated on the tomato cultivar “Zhongshu-4” (susceptible to M. incognita) in a greenhouse with a 16/8-h light/dark cycle at 21–26°C. Egg masses were extracted from tomato roots on the 42nd day after inoculation. Eggs were collected on a 25 μm sieve and placed in an incubator at 28°C for hatching second-stage juveniles (J2s). Fresh J2s were collected daily and used as inoculums for testing nematode mortality and infection assays.

Bv-DS1 was isolated from a tidal soil sample collected in Dongying, Shandong Province, China, using the 10-fold dilution method on lysogeny broth (LB) medium. The complete 16S rRNA was sequenced in BGI (Shenzhen, China) and blasted using EzBioCloud (https://www.ezbiocloud.net/). The purified strain was stored in the China centre for type culture collection (CCTCC) with the accession number CCTCC M 2021239. Genomic DNA was extracted using the blood and cell culture DNA midi Kit (Cat. No. 13343, Qiagen, United States) according to the manufacturer’s protocol. Briefly, an appropriate volume of cultured bacteria was pelleted by centrifugation at 4,000 × g for 10 min and the supernatant was discarded. The bacteria pellet was then resuspended in 3.5 ml of Buffer B1 (with RNase A) by vortexing at top speed. A stock solution of 20 μl lysozyme stock solution (100 mg/ml) and 100 μl QIAGEN Protease or QIAGEN Proteinase K was added and incubated at 37°C for at least 30 min. Next, 1.2 ml of Buffer B2 was added and mixed by vortexing for a few seconds, followed by incubation at 50°C for 30 min. Then, the sample was vortexed for 10 s at maximum speed and applied to the equilibrated QIAGEN Genomic-tip. The QIAGEN Genomic-tip was washed with 2 × 7.5 ml of Buffer QC, followed by the precipitation, purification and dissolving of DNA. DNA concentration and purity were determined using a Qubit fluorometer and NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Carlsbad, CA, United States). DNA integrity was assessed using 0.5% agarose gel electrophoresis.

Whole genome sequencing was performed using the MGISEQ-2000 platform and Oxford Nanopore Technologies (ONT) PromethION P24 device at BGI (Shenzhen, China). For the MGI sequencing library, the insert size was 350 bp with a pair-end sequencing length of 150 bp. Briefly, 1 μg of genomic DNA was randomly fragmented using a g-TUBE device (Covaris, Inc., Woburn, MA, United States) according to the manufacturer’s instructions. The DNA fragments with an average size of 200–400 bp were selected using magnetic beads. The selected fragments were 3′-adenylated through end-repair and adapter-ligation; PCR products were purified using the magnetic beads. The double-stranded PCR products were heat denatured and circularised using the splint oligo sequence. The single-strand circular DNA (ssCirDNA) was formatted as the final library and qualified using FastQC. For ONT sequencing, genomic DNA was used to construct a library using a ligation sequencing kit (SQK-LSK109) and native barcoding kit (EXP-NBD114) according to the standard 1D native barcoding protocol provided by the manufacturer (Oxford Nanopore, Oxford, UK). Briefly, 48 μl of genomic DNA was mixed with 3.5 μl NEBNext FFPE DNA repair buffer (New England BioLabs, Ipswich, MA, United States), 2 μl NEBNext FFPE DNA repair mix (NEB), 3 μl ultra II end-prep enzyme mix (NEB), and 3.5 μl ultra II end-prep reaction buffer (NEB) in a 200 μl PCR tube. The mixture was incubated at 20°C for 5 min followed by 65°C for 5 min. Next, 500 ng end-prepped samples were mixed with 2.5 μl native barcode (one barcode per sample) and 25 μl blunt/TA ligase master mix. The mixtures were incubated at 28°C for 10 min. A total of 700 ng pooled and barcoded DNA was used to perform adapter-ligation by adding 20 μl NEB next quick ligation reaction buffer (5×), 5 μl adapter mix II and 10 μl quick T4 DNA ligase. The mixture was incubated for 10 min at room temperature. The constructed library was quantified using a Qubit DNA HS assay kit in a 4.0 Fluorometer (Invitrogen, San Diego, CA, United States) and then loaded into the flow cell R9.4.1 of a PromethION P24 device (BGI-ShenZhen, China).

All the raw data were trimmed using SOAPnuke v.1.5.2 (Li D. et al., 2015). High-quality reads were assembled de novo using Megahit software (Chen et al., 2018). Assembled contigs with lengths less than 300 bp were discarded in the subsequent analysis. The prediction of coding genes (CDS) was analysed with Glimmer (version 3.02) and the annotation was done by alignment against the COG, Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. Secondary metabolites analysis was performed using antiSMASH 5.0 software (Blin et al., 2019).

The ability of Bv-DS1 to inhibit Rhizoctonia solani, Fusarium avenaceum and Fusarium graminearum were investigated using a plate confrontation method according to the description by Gao et al. (2021). Briefly, the pathogenic fungi were inoculated in the centre of the PDA plate and 100 μl Bv-DS1 (OD₆₀₀ = 1) were inoculated 2.5 cm from the centre containing pathogenic fungi, with plates not inoculated with Bv-DS1 used as a control. All plates were incubated at 27°C for several days until the pathogenic fungi on the control plates grew all over the petri dish. Then the growth diameter of the pathogen was measured.

The measurement of IAA production was conducted by using a modified quantification method based on Bric et al. (1991). Briefly, Bv-DS1 was cultured for 24 h in 1 ml of LB liquid medium. Then 10 μl of bacterial inoculums were transferred into the same medium supplemented with 100 μg mL⁻¹ of L-tryptophan (Sigma-Aldrich) followed by 7 days of incubation at 28°C on a shaking incubator (200 rpm/min). Then, bacterial cells were removed from the culture medium by centrifugation (4,000 × g, 5 min). The supernatant was then transferred with Salkowski reagent (49 ml of 35% HClO_4, 1 ml of 0.5 M FeCl₃) to an ELISA plate in a 1:1 the ratio, which was incubated at room temperature for 35 min. The absorbency was then read at 490 nm by using a multi-functional enzyme labeller (CLARIOstar Plus, BMG, Germany). The uninoculated tryptophan-containing medium mixed with Salkowski reagent was used as a blank. Three independent cultivations were used as triplicated replicates. A standard curve was generated from serial dilutions of IAA stock solution.

Bv-DS1 was cultured in 100 ml of LB liquid medium for 48 h at 28°C on a shaking incubator (200 rpm/min). The fermented bacteria were centrifuged at 2,500 × g for 10 min, and the bacterial pellet was resuspended and adjusted to a density of 1.0 × 10⁸ colony-forming units (CFUs) per millilitre with sterile water, which was used further for inoculation. Additionally, the supernatant of the Bv-DS1 strain was collected and filtered using a 0.22-μM Millipore filter. The prepared culture filtrate was used to test nematocidal efficacy in vitro.

For inoculation with Bv-DS1, the 3-week-old tomato plants were transplanted to individual pots (14 cm in height and 12 cm in diameter) filled with sterilized sand-soil medium (2:1, vol/vol) for another 3 days until the initiation of the experiment. Each transplanted plant was inoculated with Bv-DS1 by adding 20 ml of bacterial suspension into 2-cm-deep holes. Subsequently, the plants were put back in the growth chamber and used for nematode inoculation after 3 days of growth.

For nematode mortality assay, 10 μl of nematode suspension (approximately 100 J2s) and 490 μl of Bv-DS1 culture filtrate (100 and 10%, respectively) were added to each well of the 24-well culture plate (Corning, United States), and sterile water was used as the control. The plates were incubated in darkness for 24 h at room temperature, and then the number of living and dead nematodes were counted under a stereomicroscope (Olympus, Japan). J2s were considered to be dead if their body was straight and immobile after the Na₂CO₃ stimulus for 30 s (Hu et al., 2019). The experiment was carried out twice with 10 replicates. Corrected J2s mortality was calculated using the following equation:

[(mortality rate of J2s treated with Bv-DS1 - mortality rate of J2s treated using the sterile water)/(1 - mortality rate of J2s treated using sterile water)] × 100.

To investigate the effects of Bv-DS1 on the nematode-invasion ability, gall formation, and host defence, a pot experiment was conducted at different time points using four treatments: (1) roots treated with sterile water, (2) roots pre-inoculated with Bv-DS1 alone, (3) roots inoculated with nematodes alone, (4) the roots preinoculated with Bv-DS1 and then infected with nematodes. Three-week-old tomato plants were transplanted to pots filled with sand-soil medium and were grown for 6 days. Two holes were opened on the surface of pots, and plants were inoculated with 500 J2s per plant in one pot. At 3, 7, and 14 days after nematode inoculation (dai), roots were collected and used for RNA extraction and gene expression analysis. At 35 dai, plant height, fresh weight of root and shoot, and stem thickness were measured, and the disease severity was assessed by counting gall numbers on roots. Each treatment was performed with six replicates, and three independent experiments were conducted for each treatment.

To further evaluate the control efficacy of Bv-DS1 against RKN in the greenhouse condition, the soil collected from the M. incognita-infested tomato field (Shenyang, Liaoning province, China) was used in the pot experiment. The population densities of M. incognita were 1.5 nematodes per cm³ of soil. Three-week-old tomato plants were transplanted into pots (20.5 cm deep × 14 cm diameter) filled with soil media containing 20% sterilized sand and 80% diseased soil. After 3 days of transplantation, 20 ml of the Bv-DS1 suspension (10⁸ CFU/ml) was drenched into the rhizosphere of the plant in each pot; the plants were irrigated with the same volume of sterile water that served as the control. All plants were grown in the greenhouse under a completely randomized design for 42 days. Afterwards, the roots of tomato plants were collected, and fresh root weight was measured. Root galls per plant were counted and the number of egg masses was determined using 0.01% erioglaucine (Sigma, St. Louis, MO, United States) staining (Omwega et al., 1988). Finally, eggs were extracted from the roots according to the previously described method (Hussey and Barker, 1973). Each treatment had eight replicates, and the experiment was repeated twice.

The split-root system of tomato was used to evaluate the ability of Bv-DS1 to induce systemic plant resistance against M. incognita as described by Martínez-Medina et al. (2017). Three-week-old tomato plants were transferred to the split-root system by splitting the root system into two halves that were planted into two adjacent pots (14 cm in height and 12 cm in diameter) containing a sterilized sand-soil mixture (2:1, v/v) (Figure 1A). A total of three treatments were used in this experiment with eight replicates for each treatment. The treatments included (1) half of the root system being inoculated with 500 J2s of M. incognita (RKN/−) and another half of the roots with sterile water, (2) half of the roots were pre-inoculated with the Bv-DS1 suspension and then infected with nematodes (RKN + Bv-DS1/−), and (3) half of the root system was pre-inoculated with Bv-DS1 and another half was infected with nematodes (RKN/Bv-DS1). For treatments of Bv-DS1, the plants in the split-root set-up were inoculated with 20 ml of suspension of Bv-DS1 (10⁸ CFU/ml) 6 days after transplantation. The pots were placed in the greenhouse under the same condition for another 3 days.

Tomato roots were frozen in liquid nitrogen and ground to a fine powder in a pestle and mortar. RNA was extracted using an RNAprep pure plant kit (TianGen Biotech, Beijing, China) according to the manufacturer’s instructions. One microgram of total RNA was used to synthesize cDNA using FastKing gDNA dispelling RT SuperMix FastKing Kit (TianGen Biotech, Beijing, China). qRT-PCR analysis was performed in the LightCycler® 480 System with SYBR green master mix (Vazyme, Nanjing, China). The target gene primers used for qRT-PCR are listed in Supplementary Table S1. Reaction conditions were as follows: 95°C for 5 min, and then 40 two-step cycles of 10 s at 95°C and 30 s at 60°C. The relative expression levels of the defence-related tomato genes and the actin gene from M. incognita were normalized and calculated using the reference gene expression of SIEF1α using the 2^-∆∆Ct method.

Data analysis was performed using SPSS version 17.0 software (SPSS Inc., Chicago, United States). The statistically significant differences were analysed using one-way ANOVA (multiple comparisons) or Student’s t-test (unpaired comparisons), as shown in the figure legends. The error bars in the figures indicated the standard error (SE) of means, and the significance level was set at p < 0.05.

Analysing the complete 16S rRNA sequence (1,471 bp) of the stain using EzBioCloud revealed that it showed 100% similarity with B. velezensis CR-502. Phylogenetic analysis indicated that it grouped with the B. velezensis strain CR-502 T (AY603658), thus confirming its classification as B. velezensis (Supplementary Figure 1). Further analysis of genome characteristics revealed that the genome of Bv-DS1 comprised a circular chromosome of 4.73 Mb (Figure 2A), and was deposited in NCBI with the accession number CP102866. The chromosome of Bv-DS1 included 4,007,438 bp, with an average GC-content of 46.43%, 3,977 protein-coding genes (CDS), 86 tRNAs, and 27 rRNAs. A total of 4,334 (82.2%) CDSs were classified into Cluster of Orthologous Groups of proteins (COG) families composed of 25 categories (Figure 2B). Among the categories, amino acid transport and metabolism (306 genes), transcription (288 genes), and cell wall/membrane/envelope biogenesis (207 genes) were the top three functional categories. However, there was a high proportion of function unknown genes (206), and general function prediction-only genes (330 genes) were poorly characterized (Supplementary Table S2).

Using the antiSMASH genome analysis tool, the detection of secondary metabolite clusters of Bv-DS1 were detected (Supplementary Table S3). Three trans ATPKS, two NRPS, and one other (bacilysin) cluster showed 100% similarity to the known biosynthetic gene clusters. Several clusters related to surfactin, aurantinin B/aurantinin C/aurantinin D, and butirosin A/butirosin B saccharides were also detected in the Bv-DS1 genome (Supplementary Table S3).

As Bv-DS1 promoted the growth of tomatoes plants, we hypothesized that it might be also able to produce IAA. Bv-DS1 produced 3.07 μg mL⁻¹ IAA and showed antifungal activity against three soybean pathogens (Rhizoctonia solani, Fusarium avenaceum, and Fusarium graminearum) that cause root rot disease, with the inhibition zones of 1.04 ± 0.18 cm, 1.04 ± 0.15 cm and 1.70 ± 0.20 cm, respectively (Supplementary Figure 2). The nematicidal activity of Bv-DS1 was assessed in a 24-cell plate by analysing the mortality rates of M. incognita J2s after treatment with Bv-DS1 culture. After incubation for 24 h, the 1 × Bv-DS1 and 5 × filtrates resulted in 71.62 and 43.16% corrected J2s mortality rates of M. incognita, respectively (Figure 1A). The microscopic observation indicated that most J2s were straight and immobile in the filtrate culture of Bv-DS1 after 48 h treatment. In comparison, the untreated J2s displayed the normal ‘S’ bend shape and were much more active (Figure 1B). These results demonstrated that Bv-DS1 metabolites had nematicidal activity against M. incognita in vitro.

The efficacy of Bv-DS1 against M. incognita in tomatoes was evaluated in the pot assay. After 5 weeks of transplantation, Bv-DS1 treatments caused a significant increase in the plant height and root and shoot weight of tomato plants, inoculated or non-inoculated with nematodes, suggesting that Bv-DS1 had a positive effect on plant growth (Figure 3A; Table 1). The expression of the actin gene of M. incognita was determined in tomato roots at 3 and 7 dai to investigate the effect of Bv-DS1 on M. incognita infection (Martínez-Medina et al., 2017). Although the actin gene of M. incognita showed low expression levels in both Bv-DS1-treated and untreated plants at 3 dai, actin gene expression in Bv-DS1-treated roots was significantly lower than that in the non-inoculation roots (Figure 3B). At 7 dai, pre-inoculation with Bv-DS1 led to a 1.5-fold reduction in expression levels of actin compared to the roots without Bv-DS1 treatments (Figure 3B). These results showed that the application of Bv-DS1 inhibited early infection of M. incognita. Additionally, the number of root galls on tomato roots pre-inoculated with Bv-DS1 (100 ± 11 per plant) was significantly lower than that of the non-inoculated control roots (200 ± 11 per plant). These results indicated that Bv-DS1 had plant growth promoting potential and could efficiently control M. incognita (Figure 3C).

The control efficacy of Bv-DS1 against RKN was tested further in the soil collected from the M. incognita-infested tomato field. Irrigation of Bv-DS1 into the rhizosphere significantly increased the fresh weight of the tomato plants by 35.5% compared to the control group at 42 days (Figure 4A). The number of galls and egg masses were lower on the tomato roots after treatment with Bv-DS1 (Figure 4B), which decreased by 29.3 and 33.8%, respectively (Figure 4C). Eggs per egg mass from the root system in the soil of Bv-DS1 drenching was also markedly lower than those collected from the non-inoculated plants (Figure 4D). These results suggested that Bv-DS1 could enhance the resistance of tomatoes to suppress M. incognita reproduction.

To assess whether the Bv-DS1-mediated plant resistance to RKN occurred in the systemic root tissue of tomato, a split-root system of tomato (Martínez-Medina et al., 2017) was used (Figure 5A). Compared to the control roots (only inoculated with RKN), pre-inoculation with Bv-DS1 caused a reduction in the galling of the local root system (Bv + RKN/−) (Figure 5B). The number of galls and egg masses in the local root system of the Bv-DS1-treated plants significantly decreased by 48.42 and 64.81%, respectively (Figures 5C,D). Although there was a slight reduction in the number of galls and egg masses in the systematic root tissue after Bv-DS1 treatments, no statistical significance was found in the Bv-DS1-induced systemic protective effects when compared to the control group (Figures 5C,D). These results indicated that Bv-DS1 could not elicit systemic resistance to RKN in tomatoes.

SA and JA are two important plant hormones and play crucial roles in plant defence response to nematode infection. To examine whether SA or JA-dependent signalling contributed to the Bv-DS1-mediated tomato resistance to RKN, we analysed the detailed transcript abundance of the SA and JA marker genes in Bv-DS1-preinoculated tomato roots under RKN stress (Figure 6). The expression of JA-related genes LOX D and MC in the roots of the RKN-untreated plants was significantly upregulated via Bv-DS1 preinoculation. RKN infection also resulted in the upregulation of transcript levels of MC at 7 and 14 dai. No significant effect of Bv-DS1 on the expression of MC in nematode-infected roots was found at 3, 7, and 14 dai. However, co-inoculation of RKN and Bv-DS1 caused a transiently significant downregulation of LOX D expression at 7 dai (Figure 6). Expression levels of SA-responsive genes PAL2 and PR in the roots of Bv-DS1 pretreated plants were similar to that of the non-inoculated tomato roots at 3 and 7 dai. Similar to the changes in JA marker genes, PAL2 and PR transcripts were significantly upregulated via Bv-DS1 preinoculation or RKN infection alone, but this activation of RKN-induced PAL2 and PR expression was not observed in Bv-DS1 preinoculated roots at 14 dai.

The SA-deficient transgenic NahG tomato line, the JA-deficient mutant spr2 and their corresponding background wild-type tomato lines ‘Castlemart’ and ‘Moneymaker’ were used to further assess the roles of SA and JA pathways in the biocontrol effects of Bv-DS1 in tomatoes against RKN. A significant reduction in root galls was observed in both Bv-DS1-treated ‘Castlemart’ and JA-deficient mutant spr2 when compared to the Bv-DS1 non-inoculation roots. Similarly, Bv-DS1 preinoculation also resulted in a reduction in the number of root galls in both the NahG tomato line and the wild-type ‘Moneymaker’ at 21 dai (Supplementary Figure 3). These findings indicated that Bv-DS1-induced tomato resistance against RKN was not dependent on the SA or JA pathways.

Tonoplast intrinsic proteins (TIPs), localised in vacuoles, play a key role in plant defences against PPNs through the regulation of water and ion transport (Baranowski et al., 2019). Therefore, three TIP genes (TIP1.1, TIP1.2, TIP1.3), which displayed significant downregulation in the RKN-infected susceptible tomato roots (Shukla et al., 2018), were selected to study the effects of Bv-DS1 on their expression in the RKN-inoculated tomato roots. The expression of TIP1.1 and TIP1.2 was significantly upregulated by Bv-DS1 at 24 and 72 h, respectively. TIP1.3 transcript levels reached the peak at 24 h and then declined at 72 h but were still higher than that of untreated control roots (Figure 7A). The expression levels of three TIP genes were significantly downregulated at 3 dai, this suppression of RKN-induced TIP1.1 and TIP1.3 expression was alleviated in tomato roots through Bv-DS1 preinoculation (Figure 7B).