The Phytophthora drechsleri Gcbs-5 strain was isolated from gomchwi in greenhouse fields in Geochang in Korea following Meszka and Michalecka (2016). It was then stored on a potato dextrose agar (PDA; Difco, Detroit, MI, USA) slant in the dark at 15°C (Kwon et al., 1999). Identification of P. drechsleri was done by comparing its cultural and morphological characteristics with descriptions published in the literature and by phylogenetic analysis (Rural Development Administration, National Institute of Agricultural Science and Technology, Plant Pathology Division, 2000).

Zoospores probably function as the primary inoculum for infecting gom-chwi plants (Biles et al., 1995), however, a major problem is production of the large quantities of similar-aged zoospores of P. drechsleri required for inoculation (Mansoori and Banihashemi, 1982) according to existing methods (Barash et al., 1965; Cother and Griffin 1973; Mehrotra 1970). Thus, the mycelium of P. drechsleri were used as a pathogen inoculum for seedling assay and tests on potted plants in the glasshouse. Forty grams of barley was added to 45 ml of distilled water in a 1 l Erlenmeyer flask and then it was autoclaved twice for 20 min at 121°C. The flask was inoculated with five mycelial plugs (5 mm diameter) from a 5-day-old culture of P. drechsleri, that was incubated at 25°C. The inoculated barley medium was shaken daily to break up particles so that the pathogen and medium were well mixed. For pathogen inoculation, a granule of P. drechsleri infected barley was attached to the upper part of gom-chwi roots (Pieterse et al., 1996).

Bacteria were isolated from the rhizosphere soil collected from seven kinds of plants: gom-chwies, strawberries, celeries, kales, red beets, chicories, and red lettuces from Damyang, Iksan, Milyang and Sunchang regions in Korea. A threegram soil sample was added to 27 ml of sterilized distilled water and shaken at 180 rpm and 28°C, for 30 min. Subsequently, 1 ml suspension was taken and diluted to 9 ml of sterilized distilled water, 5 times in a row, and 1 ml of each final suspension was taken and spread onto tryptic soybean agar (TSA; Difco), 1/10 TSA and nutrient agar (Difco) medium. After incubation at 28°C for 48 h, single colonies with different colors and shapes were isolated and stored at -70°C in 20% glycerol.

Inhibition of mycelial growth of P. drechsleri in the bacterial isolates was tested by dual culture assay. The bacterial isolates were equidistantly streaked on the margins of the PDA plates and incubated at 28°C for 24 h. Sterile distilled water was used for the control. The agar disc with 5 mm of P. drechsleri was placed at the center of the PDA plate and incubated at 25°C for 5 days. The growth inhibition length of mycelium was measured in order to screen the antagonistic bacterial strains against P. drechsleri, using the method of Idris et al. (2007).

Forty-nine bacterial strains, resulting from dual culture assays, were evaluated for the suppression of Phytophthora root rot using gom-chwi seedlings. For the seed dipping treatment, 10⁸ cfu/ml (OD₆₀₀ = 0.25) of bacterial suspension in 0.1% peptone water (Becton Dickinson Biosciences, San Jose, CA, USA) was prepared. The seeds (cv. Aram) were dipped in a bacterial suspension for 3 h. For the pathogen-treated control (control) and pathogen non-treated control (mock), seeds were dipped in 0.1% peptone water without any bacterial treatment. The gomchwi seeds were sown on a 32-hole tray filled with soil (Green Soil, Seoul Bio, Eumseong, Korea). A set of 16 individual gom-chwi plants was used for each treatment. To introduce the pathogen inoculum, one granule of barley infected with P. drechsleri, as described above, was inoculated on the upper part of the root of a 5-week-old gom-chwi after the soil was slightly removed. Pathogen non-treated control plants were mock-inoculated with a granule of noninfected barley (Maleki et al., 2011; Pieterse et al., 1996). All plants were irrigated regularly. The severity of diseased plants was evaluated 7 days after pathogen inoculation and the onset of symptoms. Disease severity (modified from Irabor and Mmbaga, 2017) was evaluated on a scale of 0-3 (Supplementary Table 1): 0 = no visible symptoms; 1 = slightly diseased, 20% of total leaves were wilted with brownish lesions beginning to appear on the stem; 2 = stem lesions extending to cotyledons or petioles and 50% of total leaves were wilted; and 3 = entire plant diseased or dead.

Six bacterial strains from the seedling assays with seed dipping treatment were evaluated using gom-chwi plants using a soil drenching treatment. The bacterial strains were incubated in the tryptic soybean broth medium (TSB; Difco) at 28°C for 24 h. The culture solutions of bacterial strains were adjusted to a concentration of 10⁸ cfu/ml (OD₆₀₀ = 0.25). Five-week-old gom-chwi seedlings growing in a 32-hole tray were drenched once with 15 ml bacterial culture solution, and a set of 32 individual gom-chwi plants was used for each treatment. The control and mock-control were treated with 15 ml of TSB medium. Pathogen inoculum was conducted as described above and, disease severity were also evaluated as described above. All experiment were repeated three times. The control efficacies indicate the differences in the disease incidence rates, as a percentage, of between bacterial treatments and untreated control, which was calculated by the method of Tsuda et al. (2016).

The total genomic DNA of the ObRS-5 strain was extracted using a Genomic DNA Prep kit (Nanohelix, Daejeon, Korea) following the user manual. A 1,356-bp of 16S rRNA gene fragment was amplified from genomic DNA of strain ObRS-5 using the universal primers, 27F (forward) 5ʹ-GAG TTT GMT CCT GGC TCA G-3ʹ and 1492R (reverse) 5ʹ-ACG GYT ACC TTG TTA CGA CTT-3ʹ (Weisbrug et al., 1991). Amplification of the 16S rRNA region was performed in a total volume of 25 μl containing 50 ng of DNA template, 0.25 μM of each primer, 0.25 μM of dNTP (Inclone, Daejeon, Korea), 2.5 μl of 10× buffer (Inclone) and 1.25 U of Taq DNA polymerase (Inclone), and 1.25 μl of 1.10 g/ml of dimethyl sulphoxide (Sigma-Aldrich, St. Louis, MO, USA). The PCR mixture was first heated at 95°C for 5 min, and 30 cycles were then performed at different temperatures: 95°C for 1 min, 55°C for 1 min, and 72°C for 1 min 30 s. The reaction was finally incubated at 72°C for 10 min. The fragment size of the amplified PCR products was confirmed by electrophoresis analysis, and afterward, sequencing was conducted by Genotech Co. (Daejeon, Korea). The raw sequences were aligned for a phylogenetic dendrogram, which was constructed using a neighbor-joining algorithm using MEGA software (ver. 6.06) and 1,000 replicates of bootstrapping (Felsenstein, 1985). The working reference sequences were downloaded using the EzBi°Cloud server at http://www.ezbiocloud.net/.

The culture solution of the ObRS-5 strain was adjusted to 1.0 × 10⁸ cfu/ml (OD₆₀₀ = 0.25) and centrifuged at 8,000 rpm for 10 min. The supernatant was then filter-sterilized using a 0.22 μm syringe filter (Millipore, Billerica, MA, USA) for ‘(filtered) filtrate’ or was autoclaved at 121°C for 15 min for ‘sterilized filtrate’. The harvested cells were resuspended in 0.1% peptone water at a concentration of 1.0 × 10⁸ cfu/ml (OD₆₀₀ = 0.25). Mycelial discs of P. drechsleri from the peripheral growth of a 5-day-old culture on PDA were transferred to Petri dishes, containing 30 ml of 10% (v/v) clarified V8 medium with 0.3% CaCO₃, which were incubated at 25°C for 4 days (Lee et al., 2001). The medium was decanted and replaced by 30 ml of bacterial cell suspension, culture solution, and filtrate of E. asburiae ObRS-5, with the sterile soil extracts at a ratio of 1:6 (v/v) (De Cock and Lévesque, 2004). Soil extracts were prepared as described by Mansoori and Banihashemi (1982). The control was treated with only sterile soil extracts, and the medium-treated control was treated with TSB and sterile soil extracts at a ratio of 1:6 (v/v).

All plates were light-irradiated at 24 h per day at 25°C for 2 days. Microscopic observations (M205A, Leica, Jena, Germany) were conducted to count the number of zoosporangia formed (zoosporangia/5 mm2) from five random areas of the fungal growth, from each mycelial disc, because zoosporangium of P. drechsleri is non-caducous (Singh et al., 1992). Thus, the total number of zoospores was also investigated. In order to release the zoospores, all plates were treated at -20°C for 25 min and treated at room temperature for 3 h. The mycelium was filtered with sterile gauze and the released zoospores were counted using a hemocytometer (Incyto, Cheonan, Korea) under an optical microscope (DM2500, Leica).

A zoospores suspension of P. drechsleri was prepared at a concentration of 1.0 × 10⁴ zoospores/ml as described above. A prepared cell suspension, filtrate, or culture solution of ObRS-5 strain was mixed with the sterile soil extract at a ratio of 1:1 (v/v) in a 96-well cell culture plate (SPL, Pocheon, Korea). Zoospore germination was observed under a microscope at 2-h intervals during incubation at 25°C. The result of the cell suspension treatment was compared to the soil extract-treated control and the result of the remaining treatments were compared to the medium treated control.

Gomchwi were sown in a 32-hole tray, they were grown for 4 weeks, and then gom-chwi plants were transplanted into plastic pots (10-cm-diameter) with soil. A set of 30 individual 12-week-old gom-chwi plants was used for tests under glasshouse conditions. Each plant was drenched once with 50 ml of cell suspension, filtrate, sterilized filtrate, culture solution of E. asburiae ObRS-5 or TSB medium for control and mock-control treatments. Pathogen inoculum was treated with two granules of barley infected with P. drechsleri as described above. Disease severity were also evaluated as described above.

Data were statistically verified whether it follows a normal distribution or not by Levene test, and analyzed using analysis of variance (ANOVA) in R software (ver. 3.4.1), and the significance of each value was determined using Duncan’s multiple range test (P < 0.05).

A total of 172 isolates were finally obtained from rhizosphere soils from diverse crops. Among these bacterial isolates, 26 isolates had 10-20% and 144 isolates had less than 10% inhibition effect on mycelial growth in P. drechsleri (Table 1, Supplementary Fig. 1). In addition, only 2 isolates showed more than 20% mycelial inhibition rate, which was 1.2% of the total bacterial isolates. Thus, the evaluation with the seedling assay was carried out on 49 bacterial strains (28.5% of isolates) having a varied range of mycelial growth inhibition abilities.

Among the 49 antagonistic strains, in the seed dipping treatment of six bacterial strains, RlR9, ObRS-5, ObRS-6, ObRS-10, ObRS-15, and HRS-20 were the most efficient in reducing disease severity in gom-chwi seedlings compared to the bacterially untreated control (Supplementary Table 2). The value of disease control efficacy (%) of RlR9, ObRS-5, ObRS-6, ObRS-10, ObRS-15, and HRS-20 was 76.5, 76.5, 70.6, 70.6, 64.7, and 58.8, respectively. Whereas 22 strains showed a mildly suppressive effect with no significant differences compared to the control and 21 strains had no effect (0%) (Supplementary Table 2). Thus, six bacterial strains were selected as potentially effective antagonists for further tests.

To assess whether these six bacterial strains have suppressive effects in sterilized soils or not, we performed to shoot plate assay (Supplementary Fig. 2A). In the shoot plate assay, seed dipping treatment of RlR9, ObRS-5, ObRS-6, ObRS-10, ObRS-15, and HRS-20 strains were showed means of infected rate at 15.6%, 12.5%, 31.3%, 34.4%, 34.4%, and 15.6%, respectively, and those of ObRS-6, ObRS-10, and ObRS-15 strains have no suppressive effects compared to water-treated control. Whereas RlR9, ObRS-5, and HRS-20 showed control efficacy at 29.2%, 45.8%, and 25%, respectively (Supplementary Fig. 2B).

In the seedling assay with soil drenching treatment, three strains, ObRS-5, ObRS-6, and ObRS-15, showed significant control efficacies of 90.5%, 77.8%, and 66.7%, respectively (Fig. 1, Supplementary Fig. 3). Thus, among the six antagonistic strains, the ObRS-5 strain was selected as potentially the most effective antagonist for further testing.

Molecular phylogenetic analysis of the 16S rRNA sequences of ObRS-5 strain showed 99.9% homology with Enterobacter asburiae JCM 6052, and the bootstrap value was 86% indicating that it belongs to the subcluster of E. asburiae (Fig. 2). Consequently, the ObRS-5 strain was finally identified as Enterobacter asburiae ObRS-5. The nucleotide sequences of ObRS-5 strain was submitted to Genbank as the accession number MT020371.

Zoosporangia formation was reduced by treatment with cell suspension (87.4%), filtrate (97.0%), sterile filtrate (98.5%) and culture solution (100%) of E. asburiae ObRS-5 (Fig. 3). As a result of counting the number of zoospores, the inhibition rate compared to the control was highest for the sterile filtrate (94.6%), followed by the filtrate (91.7%), culture solution (87.1%) and cell suspension (31.5%) of E. asburiae ObRS-5 (Fig. 3).

Meanwhile, in the zoospore germination, treatment of cell suspension, filtrate, sterile filtrate, and culture solution of E. asburiae ObRS-5 showed a germination rate of 33.3%, 32.4%, 93.6%, and 9.6%, respectively. The germination rate of sterile filtrate did not show significant differences in either medium treated control (90.6%) or soil extract-treated control. There was also no statistical difference between the medium treated control and soil extract-treated control. As a result, the cell suspension, filtrate, and culture solution reduced zoospore germination at 66.7%, 90.4%, and 67.6%, respectively (Fig. 4). In particular, E. asburiae ObRS-5 culture solution effectively suppressed zoospore germination. Therefore, E. asburiae ObRS-5 or its products might be involved in the inhibition of zoosporangia formation of P. drechsleri.

The E. asburiae ObRS-5 strain inhibited P. drechsleri infection in potted plants in the glasshouse. It was significantly antagonistic, and the control efficacies of gom-chwi plants were 80% (culture solution), 63.2% (cell suspension), 55.6% (filtrate) (Fig. 5). Gom-chwi plants treated with sterilized filtrate of E. asburiae ObRS-5 showed a disease severity of 1.08, of which the pathogen-treated controls were 1.50 (medium) and 1.58 (water), and the results revealed that there was not significantly suppression effect, which is similar to the result of the in vitro experiment.