A B. velezensis strain GH1-13 was isolated from a reclaimed paddy field on Wando Island in Korea (Kim et al., 2016). To investigate its antifungal activities against turf pathogens, fungal and oomycete strains causing or potentially causing brown patch [R. solani AG-1(IA) (KACC 40103), R. solani AG-1(IB) (KACC 40109), R. solani AG-4 (KACC 40143), R. solani AG2-2(IIIB) (KACC 40151)], large patch [R. solani AG2-2(IV) (KACC 40132)], yellow patch [R. cerealis (KACC 40154)], Pythium blight [Pythium ultimum (KACC 40705)] were obtained from the Korean Agricultural Culture Collection (KACC), National Institute of Agricultural Sciences, Rural Development Administration, Jeonju, Korea. In addition, one strain (CMML20-28) of the dollar spot pathogen Clarireedia jacksonii isolated from creeping bentgrass leaf blade and identified by multiple markers (unpublished data) was also used in this study. This bacterial strain was cultured in tryptic soy broth (TSB, Difco, Sparks, MD, USA) and tryptic soy agar (TSA), and the fungal and oomycete strains were cultured on potato dextrose agar (PDA, Difco, Sparks, MD, USA). The cultured bacterial, fungal and oomycete strains were maintained in 20% glycerol stock solution at − 80°C and on TSA and PDA plates at 4 °C prior to use.

Two fungicides, azoxystrobin (Otiva, Syngenta, containing 21.7% of active ingredient) and fluxapyroxad (Cardis, Nonghyup chemical, containing 15.3% of active ingredient), were used in the study. Stock solutions (10,000 μg ml^-1) of azoxystrobin and fluxapyroxad were prepared and stored at 4°C before use.

Antifungal activity of B. velezensis GH1-13 against stains of five different anastomosis groups [AG-1(IA), AG-1(IB), AG-4, AG2-2(IIIB), and AG2-2(IV)] of R. solani, R. cerealis, Pythium ultimum, and C. jacksonii was tested to investigate the potential for the bacterium to control turf pathogens. At first, a single colony of B. velezensis GH1-13 was inoculated and grown in 5 ml of TSB overnight at 30°C at 150 rpm in a shaking incubator. Then, 150 μl of the bacterial culture was transferred to 5 ml new TSB media. After growing the bacterial culture for 6 h, 10 μl of the culture (optical density (OD) value = 0.4; 1×10⁷cfu ml^-1) was inoculated on three positions 20 to 25 mm from the midpoint where the fungal agar plug was to be placed. Eight strains were grown on PDA for 4-5 days at 25°C, and one 5mm agar plug taken from the edge of the actively growing colonies of each isolate was placed at the center of each of the three replicate PDA plates of each isolate one day after the bacterial inoculation. Pathogens without bacteria served as the control. All plates were incubated at 25°C in darkness. The colony diameter of each pathogen in the absence or presence of GH1-13 was observed and was measured 2-14 days after inoculation. The inhibition rate was calculated using the following formula (Ji et al., 2013):

A1 is the average diameter of colonies from each pathogen and A2 is the average diameter of colonies in the presence of GH1-13. The experiment was performed twice with three replicates.

To evaluate the antifungal efficacy of B. velezensis GH1-13 against brown pathogen, the strain KACC 40151 of R. solani AG2-2(IIIB) was co-cultivated with the strain GH1-13 on PDA media at 25°C for 8 days. Mycelia of R. solani AG2-2(IIIB) that were adjacent to the bacterium or growing on the opposite sides from the bacterium were placed in 10 μl of neutral red (0.1 μg ml^-1, DaeJung) or Evans blue (0.5 μg ml^-1, Alfa Aesar) on separate glass slides. The slides were incubated for 5 min at room temperature and washed 3 times with sterilized ddH₂O. The stained mycelia and the morphology of mycelia were observed, and microscopic images of mycelia were photographed under a light microscope (Olympus, Tokyo, Japan). The experiments were conducted twice, and more than 50 images of mycelia per treatment were taken.

Compatibility tests of fungicides with GH1-13 were conducted to investigate the effect of fungicides commonly used in turfgrass on B. velezensis GH1-13. The strain GH1-13 was grown in nutrient broth (NB, Difco, Sparks, MD, USA) medium on a shaking incubator at 180 rpm at 28°C for 24 hrs. Then, 150 μl of bacterial culture was added to the flasks containing 20 ml of NB medium with and without fungicides (20 and 100 μg ml^-1 of azoxystrobin and fluxapyroxad). After incubation at 180 rpm and 28°C for 24 hrs, one ml of the culture from each flask was serially diluted at 10-fold intervals in NB, and the 10^-4 fold-diluted culture was plated onto nutrient agar (NA) medium. The number of individual bacterial colonies on each plate was counted after being cultured for 24 hrs at 28°C. The experiment was conducted three times with three replications.

The efficacy of B. velezensis GH1-13 and fungicide (azoxystrobin) against the brown patch pathogen was tested on creeping bentgrass. The bentgrass was grown in vinyl pots containing 120g soil in a growth room at 23 ± 5 ˚C for 3 weeks. The strain R. solani AG2-2 (IIIB) KACC 40151 was grown in PDB for 3-4 days, and mycelia were ground for inoculation. Four different treatments were prepared as follows: GH1-13 suspension (1×10⁶cfu ml^-1), azoxystrobin (20 μg ml^-1), GH1-13 suspension (1×10⁶cfu ml^-1) + 50 percentage of azoxystrobin (10 μg ml^-1), and 50 percentage of azoxystrobin (10 μg ml^-1). Sterilized water was used as an untreated control. After applications of the four treatments, the ground mycelia were inoculated on bentgrass. Disease severity (using a 0 to 7 scale; 0 = healthy plant, 1 = less than 15% infection, 2 = 15-30% infection, 3 = 30-45% infection, 4 = 45-60% infection, 5 = 60-75% infection, 6 = 75-90% infection, 7 = dead plant) was rated 14 days after inoculation. The experiment was conducted twice with three replicates. The disease severity index (%) was calculated (Chiang et al., 2017) as follows:

Here, DSI indicates disease severity index, n = the number of turfgrass pots in each scale, and N = total number of pots tested.

Pan-genome analysis was conducted by the BPGA program (Chaudhari et al., 2016; Kim et al., 2017) to illustrate the genomic features specific to strain GH1-13. A total of 364 genome sequences of Bacillus strains were derived from the nucleotide database of NCBI. Among them, 269 genome sequences were from B. velezensis; 95 from other species of Bacillus. All genes of these strains were analyzed by the NJ method in the pan-genome analysis pipeline with a 50% cut-off for protein sequence identity. Orthologs protein coding sequences were used to identify core conserved genes and strain-specific genes.

To detect the strain GH1-13 specifically, two unique genes from plasmid and chromosomal DNA sequences of GH1-13 were selected and used for SYBR Green qPCR and TaqMan qPCR analysis. RNA of GH1-13 was extracted using an RNAiso plus kit (Takara, Shiga, Japan), and reverse transcription of RNA into cDNA was conducted by a PrimeScript 1st Strand cDNA synthesis kit (Takara, Shiga, Japan). For specific amplification of the plasmid DNA, specific primers (F_uniqueP/R_uniqueP) were designed and analyzed for specificity by BLAST. For amplification of the unique gene from chromosomal DNA, specific primers (F_uniqueC/R_uniqueC and Ft_uniqueC/Rt_uniqueC) and a TaqMan probe were designed. The sequences of primers and probes are listed in Table 1 . To test the specificity of qPCR, genomic DNA of different species (strains) including B. velezensis strains (GH1-13, JC-14 and TSA34-9), B. subtilis JC-15, Escherichia coli, and R. solani were used with the same DNA concentration of 0.1 ng/µl. For the qPCR sensitivity assay, DNA of GH1-13 was tested with concentrations ranging from 1,000,000 fg/µl (1ng) to 10 fg/µl (10-fold serial dilutions). SYBR Green qPCR analysis was performed using Universal SYBR Green Supermix (Bio-Rad, California, USA). The reaction mixtures contained 1 µl of each primer and 10 µl of SYBR Green Supermix, 1µl of template DNA, and 7 µl of sterile water. The TaqMan qPCR reaction mixtures contained 1 µl of each primer and 10 µl of Universal Probes Supermix (Bio-Rad, California, USA), 1 µl of template DNA, 1 µl of the probe, and 6 µl of sterile water.

To detect the strain GH1-13 from the rhizosphere, 1-week old creeping bentgrass grown in pots were used for three treatments (GH1-13 treated, TSA34-9 treated, and non-inoculated) with three pot replicates of each treatment. In the bacterial treatments, 5 ml suspension (1×10⁶cfu ml^-1) was inoculated twice into the pots, with an interval of three days. The bentgrass pots were placed in a greenhouse under at 25°C during the assay. After three days of cultivation, approximately 500 mg of rhizosphere soil was sampled from each pot, and genomic DNA of the soil samples was extracted using a FastDNA Spin Kit for Soil (MP Bio, Santa Ana, USA). SYBR Green and TaqMan qPCR analyses were conducted to detect the unique gene of GH1-13 from chromosomal DNA of soil samples using the primer pairs F_uniqueC/R_uniqueC and Ft_uniqueC/Rt_uniqueC.

To detect the strain GH1-13 in the field, 250 ml of bacteria suspension (1×10⁶cfu ml^-1) was treated on creeping bentgrass in Chonnam National University experimental field in Naju, South Korea, with three replications in October, 2021. Control plots were treated with the same volume of distilled water. Creeping bentgrass rhizosphere soil samples (500 mg) were randomly collected from each treatment after 3 h, 6 months and 9 months after the treatment. The method of DNA extraction from the soil samples was same as described previously. For specific detection of GH1-13 chromosomal unique gene, SYBR Green qPCR analysis was performed with three replications using primer pair F_uniqueC/R_uniqueC.

Field efficacy test of GH1-13 for management of brown patch was conducted on the creeping bentgrass in the Chonnam National University experimental fields in Naju, South Korea in 2021 and 2022. The design and size of the experimental plots were followed the methods described by Sang et al. (2019). A randomized complete block design (RCBD) was used with three replications. Each plot was 40.0 × 60.0 cm with 3.0-cm buffer strips between each plot. Four different treatments were applied as follows (GH1-13 suspension (1×10⁶cfu ml^-1), azoxystrobin (20 μg ml^-1), GH1-13 suspension (1×10⁶cfu ml^-1) + 50 percentage of azoxystrobin (10 μg ml^-1), and 50 percentage of azoxystrobin [10 μg ml^-1)] and distilled water served as control. The strain R. solani AG2-2 (IIIB) KACC 40151 was also inoculated a day after the treatment of GH1-13. The biocontrol agent GH1-13 and the pathogen KACC 40151 applied three times with two weeks interval from the first application. Disease severity was calculated 2 weeks after the last treatment using 0 to 10 scale with the infection percentage of 0 = healthy plant, 1= less than 10% infection, 2 = 11-20%, 3 = 21-30%, 4 = 31-40%, 5 = 41-50%, 6 = 51-60%, 7 = 61-70%, 8 = 71-80%, 9 = 81-90%, 10 = 91-100% or dead plant. The disease severity index (%) was calculated using the aforementioned methods.

All statistical analysis was performed using ‘JMP Statistical Discovery’ software package, version 10.0 (SAS institute). One-way ANOVA was used to assess the analysis of datasets from dual culture assay, compatibility of fungicides, pot, and field assay. Fisher’s least significant difference (LSD) means comparison (P = 0.05) was performed to determine statistical significance. The statistical analysis of association between variables was compared using the pairwise Pearson’s correlation coefficient.

To assess the inhibitory ability of B. velezensis GH1-13 against turf pathogens, in vitro dual culture assays against eight fungal and oomycete strains were conducted. Strain GH1-13 demonstrated strong inhibitory activity against the mycelial growth of tested pathogenic strains. This dual culture assay allowed the relative growth of pathogenic fungal and oomycete mycelia to be quantitatively assessed. Considerable variability was observed in the inhibitory activity against all eight pathogens. Strain GH1-13 showed the best inhibitory activity against R. cerealis, reducing fungal growth by 85%. Against the other 7 strains, the bacterium GH1-13 showed inhibition in the range of 63-77%. The R. solani strains causing brown patch or large patch were inhibited by GH1-13 in a range of 70-77%, a significantly higher inhibition rate than against the strains C. jacksonii and P. ultimum ( Figure 1 ). These results indicate that the GH1-13 strain is a good candidate as a potential biocomponent for turfgrass fungal and oomycete pathogens.

Mycelia of R. solani were observed under the microscope after Evans blue and neutral red staining to investigate the viability and morphological changes of R. solani AG2-2(IIIB) by the B. velezensis GH1-13. In Evans blue stains, dead cells become blue, and in neutral red stains, viable or living cells are colored red. After 8 days of co-cultivation of the pathogen and the antagonistic bacteria, dramatic breakage of the mycelia of R. solani near the GH1-13 contact zone were observed and mycelia were stained blue, while the mycelia with no alteration in the cell membrane maintained their natural coloration or faint blue. In contrast, mycelia that were adjacent to GH1-13 did not stain red (consistent with dead cells), and the viable cells were observed to be red ( Figure 2A ). In the GH1-13 growth-inhibited mycelia of R. solani, distinct globoid cells were observed at the terminal hyphae ( Figure 2B ), while the control hyphal cells without inhibition were usually linear ( Figure 2C ).

It is essential to test the compatibility of B. velezensis GH1-13 with commonly used fungicides for their successful integration for disease inhibition. GH1-13 was compatible with concentrations of 20 and 100 μg ml^-1 of azoxystrobin and fluxapyroxad, respectively ( Figure 3A ). The number of colonies of GH1-13 on 20 and 100 μg ml^-1 of azoxystrobin and 100 μg ml^-1 of fluxapyroxad amended NA medium was not significantly different from the number on non-amended NA medium. Interestingly, the low concentration (20 μg ml^-1) of fluxapyroxad increased the number of colonies of GH1-13 significantly compared to other treatments or no treatment ( Figure 3B ).

The B. velezensis GH1-13 strain and fungicide azoxystrobin were used to control brown patch on creeping bentgrass. The control efficacy (%) of GH1-13 (1×10⁶cfu/ml) alone was 71.43± 6.3 and was similar to the efficacy (%) of 50% azoxystrobin (76.19 ± 3.0). The combined application of GH1-13 (1×10⁶cfu/ml) and 50% azoxystrobin (control efficacy: 92.86 ± 4.9) showed greater significant disease reduction than GH1-13 alone or 50% azoxystrobin and control efficacy similar to 100% azoxystrobin (95.24 ± 3.0) ( Figure 4 ).

The BPGA program was used to perform pipeline pan-genome analysis and all orthologous groups among the Bacillus strains tested were identified. In the analysis, core genes are a set of genes shared with all the tested strains. Unique genes represent the set of genes in each strain not shared with other strains. The results showed that the 364 genomes contained 2 core genes, and 73 genes were uniquely found in strain GH1-13 ( Figure 5A, B ). The pan-genome and core genome were curve fitted by an exponential decay model ( Figure 5C ). The predicted core, unique and accessory genes were further classified according to COG category, and their percentage in different functional categories was determined ( Figure 5D ). The Blastx tool was used to find potential matches and the function of those 73 unique genes of GH1-13. Among them, two unique genes (encoding hypothetical proteins) showing the most specific matches to GH1-13 were selected for further study. Of these two candidates, one (positioned at CP019040.1:1872828-1873025) originated from the chromosome and the other (positioned at CP019039.1:3715-3900) is from the plasmid in GH1-13 (BioProject: PRJNA359634). These two unique genes were thus used for further detection of strain GH1-13.

To detect the GH1-13 specifically, SYBR Green and TaqMan qPCR analyses using primer sets targeting the two unique genes were conducted with bacterial and fungal DNA samples. In a specificity assay of the chromosomal unique gene, DNA of the strain GH1-13 was specifically amplified by SYBR Green ( Figure 6A ) and TaqMan ( Figure 6B ) qPCR analysis and the assay could distinguish GH1-13 from the strains (JC-14, TSA34-9) of B. velezensis and other bacterial and fungal strains (B. subtilis, Escherichia coli and R. solani). Because the amplicon (75 bp) from TaqMan qPCR was different from the one in SYBR Green qPCR (81 bp), primer pair (Ft_uniqueC/Rt_uniqueC) designed for TaqMan was also tested by SYBR Green qPCR, which showed similar results ( Figure S1 ). The specificity of this chromosomal unique gene was also tested using cDNA templates, and the cDNA sample from strain GH1-13 showed specific amplification compared to the sample from strain B. velezensis JC-14 ( Figure S2 ). The specificity of the unique gene from plasmid DNA in GH1-13 was also assayed using SYBR Green qPCR analysis. The results indicated that this gene can be specifically amplified in strain GH1-13 using either DNA or cDNA templates when compared to strain JC-14 (B. velezensis). Meanwhile, the amplification of DNA template from GH1-13 showed 4.68 Ct values, on average lower than using the cDNA template ( Figure S3 ). The qPCR assay detected the chromosomal unique gene of GH1-13 and showed high sensitivity amplifying 1 ng to 10 fg of target DNA by SYBR Green (Figure S4A) and TaqMan (Figure S4B) analyses. The linear relationships between serial diluted genomic DNA concentrations (log transformed) and Ct values were y = –3.24x + 39.95 (R² = 0.98) in SYBR Green qPCR analysis and y = –2.94x + 38.62 (R² = 0.98) in TaqMan qPCR analysis. The amplification efficiency was 103.53% in SYBR Green qPCR and 118.84% in Taqman qPCR.

Of the two unique genes, the chromosomal unique gene was selected for specific detection of the strain GH1-13 from creeping bentgrass rhizosphere soil. Only GH1-13 treated rhizosphere soil showed specific amplification by SYBR Green ( Figure 7A ) and TaqMan ( Figure 7B ) qPCR assays with average Ct values were 21.82 and 22.04, respectively. The average Ct values of TSA34-9 treated and non-treated samples all exceeded 35. The average DNA quantities detected in GH1-13 treated rhizosphere soil samples were 427,682 fg/µl and 387,165 fg/µl in SYBR Green and TaqMan qPCR analysis, respectively. For TSA34-9 treated and non-treated samples, the DNA quantities detected were all less than 50 fg/µl ( Figures 7C, D ).

To verify the detection method in the field condition, GH1-13 was treated in creeping bentgrass field and detection was conducted using soil samples after 3 h, 6 months, and 9 months of treatment. The chromosomal unique gene region of GH1-13 was successfully amplified by SYBR qPCR in all the three post-treatment samples, while the samples from the control plots remained no amplification ( Figure 8 ). The amplification curves of GH1-13 from long-term span samples (3h, 6 months, 9 months after treatment) suggested that the bacterial strain lives stably in the rhizosphere and can survive in the winter condition.

The application of GH1-13, fungicide or the combination of GH1-13 and fungicide significantly reduced the brown patch disease in the field assays in 2021 and 2022. The disease severity in the plots treated with R. solani was significantly higher (78.79 ± 3.03%) than that of the fungicide azoxystrobin 100% (6.06 ± 6.06%) and GH1-13 (9.09 ± 5.25%) treatment along with the pathogen in 2021. The control efficacy of azoxystrobin, GH1-13, and their combined application was not significantly different. In 2022, the disease severity in the plots treated with R. solani also was significantly higher (51.52 ± 8.02%) than that of the fungicide azoxystrobin 100% (9.09 ± 5.25%) and GH1-13 (12.12 ± 3.03%) treatment along with the pathogen. The treatment of azoxystrobin (100%) and the combination of GH1-13 and azoxystrobin (50%) showed highest control efficacy ( Table 2 ).