For preculture and isolation of aliquots from submerged fermentation, modified LB-medium was used containing 10 g/L of Polypepton (Nihon Pharmaceutical, Tokyo, Japan), 5 g/L of yeast extract (Oriental Co., Tokyo), and 5 g/L of NaCl, adjusted to pH 7.0. Bacillus seed inoculum was collected from a medium containing 50 g/L of fish protein (provided by Kamaboko company, Odawara, Japan), 67 g/L of glucose, 5 g/L of KH₂PO₄, 0.5 g/L of MgSO₄∙7H₂O, 25 mg/L FeSO₄∙7H₂O, 22 mg/L MnSO₄∙7H₂O and 184 mg/L CaCl₂. Agar plate containing LB medium with additional 1.5% agar (Shimizu Shokhin Kobushiki Co., Japan) was used for the colony-forming units (cfu) count. Pathogenic fungi Rhizoctonia solani K1 was cultivated in potato dextrose broth (PDB) medium containing 200 g/L of potato infusion, 20 g/L of glucose, and 10 g/L of Polypepton and the pH was adjusted to 5.6.

The strain B. subtilis RB14 [12] was transferred from culture stock (kept at −80˚C in 10% glycerol solution) into 5 ml of modified LB medium and incubated at 37˚C with 120 strokes per minute (spm) for 16 h for preculture. 400 ml of overnight cultured B. subtilis RB14 was inoculated into 40 ml of fish protein medium in 200 ml Erlenmeyer Flasks. The flasks were then incubated for 3 days at 30˚C with 120 spm for harvesting seed treatment aliquots. For harvesting fungal pathogen, at first R. solani K1 was grown on potato dextrose agar (PDA). A small mycelia disc of R. solani K1 from PDA plate was then inoculated into 40 ml of PDB and incubated in the dark at 28˚C for 7 days.

The soil used in this study was a commercially available black soil (Nittai Co., Tokyo, Japan), containing: 10.5% total carbon and 0.6% total nitrogen. It was mixed with vermiculite in a ratio of 4:1 (w/w) and the prepared soil was placed in polypropylene bag and autoclaved for 20 min at 121˚C three times at 12 h intervals. After sterilization the soil was amended with fertilizers: N: 0.04%, P: 0.09%, K: 0.06%, Ca: 0.06%, Mg: 0.06%, Fe: 0.001% and brought to approximately 60% of the maximum water-holding capacity by the addition of sterile distilled water.

Tomato seeds were disinfected with 70% ethanol and then with 0.5% sodium hypochlorite. Then the seeds were rinsed with sterile water several times. To prepare inoculum for seed treatment, two tubes containing 30 ml of a 3-day-old bacterial culture were centrifuged at 10,000× g for 10 min at 4˚C. The density of bacterial cells in the aliquots was about 10⁸ - 10⁹. The pellets were resuspended in 3 ml of sterilized distilled water. Sterilized tomato seeds were soaked into the inoculum for 15 min and the seeds were allowed to germinate on 2% agar plate at 30˚C for 3 days in the dark.

For the determination of the bacterial survivability on the seed, 10 coated seeds were taken in a sterilized polypropylene tube and after vortexing for 1 min the tube was serially diluted with 0.85% NaCl and spread on L-agar plate. The plates were incubated at 37˚C for 16 h and the number of colonies was counted. For the determination of viable cell number in soil, 3 g of soil containing RB14 was suspended in 8 ml of 0.85% NaCl solution (pH 7.0) in a 50-ml Erlenmeyer flask and shaken at 150 spm for 15 min. The suspension was used to determine the viable cell number by using L-agar plate similarly as described above.

To investigate the biocontrol activity of the B. subtilis RB14 treated seeds, tests were performed on PDA plates. A small mat of fungal pathogen was placed at the centre then B. subtilis RB14 treated seed and control seeds were inoculated at either side of the fungal mat in multiple replicate plates. Plates were incubated in a static condition at 28˚C for 5 days. Clear zone of inhibition was observed in case of B. subtilis RB14 treated seeds but the control seed area was covered by the mycelia. Inhibition of fungal growth was assessed later by measuring the size of the inhibition zone (in mm). Growth of the test fungus was calculated in percentage basis (%) using the formula: (a1 − a2/a1) × 100, where a1 represents the fungal growth (calculated as area from diameter measurements) without the antagonist (control), and a2 is fungal growth under bacterial challenge [16] .

In order to provide evidence that the antagonist B. subtilis RB14 is effectively adhered on the seeds, scanning electron microscopy (SEM) observation was performed. B. subtilis RB14 treated seeds and non treated seed samples were rinsed with sterile distilled water to remove loosely attached cells. The samples were fixed overnight in 2% glutaraldehyde in the 50 mM phosphate buffer saline (PBS) at 70% relative humidity. Samples were rinsed three times after 24 h in 50 mM phosphate buffer for 15 min each, followed by successive 15-min dehydrations in 50%, 70% and 90% ethanol and finally in 100% ethanol three times. Samples were examined by a scanning electronic microscope Hitachi S-5200 Nano SEM (Hitachi, Tokyo, Japan) operating at 1 kV.

The sterilized soil (150 g) was placed in a plastic pot with a diameter of approximately 90 mm and a height of 80 mm. Each pot was sown with nine germinated seeds and placed in a growth chamber at 30˚C with 90% relative humidity under 16 h of light (about 6000 lx).

The soil was infested with the pathogenic fungus R. solani K1 for the positive control of infection. The mycelial mats of R. solani that formed on the surface of the PDB were homogenized by a homogenizer (ACE homogenizer, Nihonseiki, Tokyo, Japan) at 4000 rpm for 2 min in sterile water and inoculated into the soil 6 days before planting the germinated tomato seeds. The inoculum was introduced into the soil at a dosage of 1 g mycelium per 100 g of soil.

To check biocontrol activity, germinated B. subtilis RB14 treated and non-treated seeds were sown into the infected pots. For control experiment germinated non-treated seeds were sown into the pots in the presence and absence of any fungal pathogen. After 3 weeks, the percentage of diseased seedlings per pot was calculated. As well as the shoots were clipped off at the soil surface level and their length and dry weight (after overnight drying at 150˚C) were measured.

To check the release of iturin A in the soil, 3 g soil from each pot was suspended in 21 ml of solvent [acetonitrile and 3.8 mm trifluoroacetic acid (4:1 v/v)] in a 50-ml Erlenmeyer flask and kept in a shaker (140 spm) for 1 h at room temperature. Soil suspension was filtered through Whatman no. 2 filter paper (Advantec, Ltd, Tokyo, Japan), and the filtrate was dried by evaporation. The precipitate was dissolved in 2 ml of methanol and distributed to 1.5 ml Eppendorf tubes. The solution was centrifuged at 15,000× g for 2 min. The supernatant was filtered through a polytetrafluoroethylene (PTFE) filter of 0.2 mm pore size (Advantec, Ltd). The filtered solution was quantified by high performance liquid chromatography (HPLC) with ODS column (Chromolith Performance RP-18e 100 - 4.6, Merck KgaA, Darmstad, Germany). The system (LC-800 system, JASCO Co. Ltd, Tokyo, Japan) was operated at a flow rate of 2.0 ml/min and monitored at 205 nm with the eluent of acetonitrile: 10 mm ammonium acetate (35:65 v/v) for measurement of iturin A.

When B. subtilis RB14 treated tomato seeds were placed on the PDA plate containing R. solani mat, clear zone of inhibition was observed (Figure 1(A)). About 28% pathogenic inhibition was observed by a single treated seed from the average value of multiple replicate plates. From the SEM image the bacterial attachment on the seed surface was further confirmed (Figure 1(B)). From cfu count it was found that the seed treating aliquot contained mainly spores. However, during the SEM examination presence of living cells on the seed surface was noticed. After seed treatment, the cells were able to re-germinate by using the seed surface nutrients as previously observed by Islam et al. 2005 [17] . It was also observed that seed treatment does not hamper the germination rate of tomato plants. Treated tomato seeds showed 99% germination index similar to the untreated seeds.

Comparative in vitro inhibition activity of B. subtilis RB14 treated and control seeds were shown in Figure 1(A) and Figure 2. As B. subtilis RB14 produces antifungal lipopeptide antibiotic iturin A, R. solani K1 cannot grow nearby as a result a large zone of inhibition was observed on the PDA plate. On PDA plate 72% fungal growth was observed in presence of a single B. subtilis RB14 treated seed whereas control seed showed 100% fungal growth as there was no growth inhibition.

In vivo fungal growth inhibition was observed when treated seeds were sown to the soil infested with R. solani K1 (Figure 3). When control seeds were sown into the R. solani K1 infested soil, 77% of tomato seedlings suffered

from damping-off disease. The B. subtilis RB14 treated seeds showed a sudden decrease in the disease occurrence which was found to be 17%. From the photographic image of the plants (Figure 4) the disease suppression ability was clearly noticed.

B. subtilis RB14 is known for lipopeptide antibiotic iturin A production. Before seed sowing there was no iturin A in the soil whereas at the day of harvest significant amount of iturin A (about 4 mg/g dry soil) was isolated from the root region soil of the tomato plants treated with B. subtilis RB14. Reasonably iturin A was not found in the soil of the control plants.