Bacillus Cyclic Lipopeptides Iturin and Fengycin Control Rice Blast Caused by Pyricularia oryzae in Potting and Acid Sulfate Soils by Direct Antagonism and Induced Systemic Resistance

Rice monoculture in acid sulfate soils (ASSs) is affected by a wide range of abiotic and biotic constraints, including rice blast caused by Pyricularia oryzae. To progress towards a more sustainable agriculture, our research aimed to screen the biocontrol potential of indigenous Bacillus spp. against blast disease by triggering induced systemic resistance (ISR) via root application and direct antagonism. Strains belonging to the B. altitudinis and B. velezensis group could protect rice against blast disease by ISR. UPLC–MS and marker gene replacement methods were used to detect cyclic lipopeptide (CLiP) production and construct CLiPs deficient mutants of B. velezensis, respectively. Here we show that the CLiPs fengycin and iturin are both needed to elicit ISR against rice blast in potting soil and ASS conditions. The CLiPs surfactin, iturin and fengycin completely suppressed P. oryzae spore germination resulting in disease severity reduction when co-applied on rice leaves. In vitro microscopic assays revealed that iturin and fengycin inhibited the mycelial growth of the fungus P. oryzae, while surfactin had no effect. The capacity of indigenous Bacillus spp. to reduce rice blast by direct and indirect antagonism in ASS conditions provides an opportunity to explore their usage for rice blast control in the field.


Introduction
Acid sulfate soils (ASSs) account for more than 1.7% of the global cultivated land worldwide and are distributed over as much as 24 million hectares [1,2]. Most of these soils are cultivated with rice because of its tolerance to acidic conditions [3]. In Vietnam, more than 1.6 million hectares of ASSs in the Mekong Delta have been ameliorated for rice cultivation by managing submergence or drainage, applying phosphate fertilizer and lime, and using resistant cultivars [4,5]. Rice monoculture in ASSs has recently been intensified with three cropping seasons per year, instead of the single or double cropping seasons of the past decades. This leads to a higher disease pressure, resulting in a drastic reduction in rice yield [6][7][8]. A prevailing disease which ravages rice fields and reduces annual rice production, especially in Vietnam, is the foliar rice blast disease caused by Pyricularia seeds were dehusked and dipped for 20 min in 2% (w/v) sodium hypochlorite solution at room temperature. Seeds were washed several times with sterile water, air dried and pregerminated for 2 days in Petri dishes containing sterile moistened filter paper at 28 • C in dark conditions. Subsequently, two day old rice seedlings were placed under greenhouse conditions with a 12 h light photoperiod at 30 ± 4 • C for 2 days to use for the experiment. Rice plants were maintained in a rice room (photoperiod with 12 h light; 30 ± 4 • C) afterwards. Bacillus isolates used in this study are listed in Table 1. The Bacillus isolates RHF4.1-25, RHF4.1-26, RHF3.1-20 and RHF2.1-7 originate from the rhizosphere of rice plants grown in ASSs in Vietnam. B. velezensis GA1 (formerly B. subtilis GA1) is a wellstudied biocontrol agent that was originally isolated from strawberry fruit [35]. All Bacillus isolates were cultured on Luria-Bertani (LB) agar at 28 • C for 24 h. Broth cultures of Bacillus strains were cultured in LB broth at 150 revolutions per minute (rpm) and 28 • C for 24 h. P. oryzae VT5M1 [12] was maintained on complete medium (CM) plates [36] at 28 • C for 5-8 days. Wild type: srf + , itu + , fen + [35] GA1∆srfaA Surfactin mutant: srf − , itu + , fen + [37] GA1∆ituA Iturin mutant: itu − , srf + , fen + [37] GA1∆fenA Fengycin mutant: fen − , itu + , srf + , [37] GA1∆srfaA-ituA Surfactin and iturin mutant: srf − , itu − , fen + This study GA1∆srfaA-fenA Surfactin and fengycin mutant: srf − , fen − , itu + This study GA1∆fenA-ituA Iturin and fengycin mutant: itu − , fen − , srf + This study Pyricularia oryzae VT5M1 Rice blast pathogen from Vietnam [12] Abbreviations: ASS: acid sulfate soil; Srf: surfactin; Itu: iturin; Fen: Fengycin.

Construction of Bacillus velezensis GA1 Mutants
Bacillus velezensis GA1 mutants deficient in the production of a single CLiP were constructed by marker gene replacement, as previously described [37]. To construct CLiP double mutants, a recombinant fragment containing a phleomycin cassette flanked by 1000 bp of the upstream region and 1000 bp of the downstream region of the targeted gene was generated by overlap PCR with specific primers (Table A1) (see Appendix A) [37]. The recombinant fragment was introduced into B. velezensis GA1 derivative mutants (GA1∆srfaA, GA1∆ituA, GA1∆fenA) by inducing natural competence, as previously described [37]. Gene replacement was selected by phleomycin resistance on LB medium. Gene deletions were confirmed by PCR analysis realized with the corresponding UpF and DwR specific primers (Table A1).

Preparation of Cell-Free Supernatants and UPLC-MS for Detection of CLiPs
To obtain cell-free culture filtrates of each isolate, a colony was transferred into 5 mL LB broth and incubated at 28 • C, 150 rpm for 24 h, then cultures were centrifuged at 13,000× g for 10 min. Subsequently, supernatants were filter sterilized through a Millipore filter (Millex-GV, 0.22 µm) to exclude bacterial cells. Cell free supernatants were collected for UPLC-MS analysis to detect and quantify CLiPs, as previous described [37], and utilized for direct antagonism assays in this study.

Potential of Indigenous Bacillus Strains to Trigger ISR against Rice Blast Caused by P. oryzae in Potting and Acid Sulfate Soil
Experiments performed to screen the ISR capacity of indigenous Bacillus spp. against rice blast caused by P. oryzae VT5M1, in both potting and ASS conditions, were conducted as described previously [24,25]. Briefly, for the preparation of bacterial inoculum, 15 mL of Bacillus seed culture was grown in 50 mL LB broth for 24 h at 28 • C. The amount of bacterial cell suspension was applied based on 600 g of sterile potting soil (Structural; Snebbout, Kaprijke, Belgium) or 800 g of sterile ASSs (a natural soil taken from representative ASSs in Kien Giang province, Vietnam), and adjusted to a final density of 1 × 10 7 CFU g −1 .
S-methyl 1,2,3-benzothiadiazole-7-carbothioate (BTH) (Syngenta Crop Protection, Brussels, Belgium), a salicylic acid analogue that is known to induce resistance against rice blast [25], was included as a positive control at a concentration of 25 µM. BTH (stock solution of 50 mM) and cell suspensions were diluted in 100 mL of water and mixed with 600 g of potting soils or 800 g of ASS for 4 min. Healthy and diseased control treatments received the same amount of tap water. Roots of rice seedlings were inoculated with the standardized bacterial inoculum by soaking for 10 min before planting them in plastic trays (23 cm × 16 cm × 6 cm) containing inoculated potting soil or acid sulfate soil. Then, bacteria were applied as soil drench 3 days before inoculation with P. oryzae VT5M1. Plants were watered every two days and trays were weekly supplied with 200 mL of nutrient solution [FeSO 4 ; 2 g/L and (NH 4 ) 2 SO 4 ; 1 g/L]. Pathogen inoculation was performed as described previously [25]. Spores of 8 day old P. oryzae VT5M1 [12] were dispersed in 0.5% (w/v) gelatin to obtain a final concentration of 5 × 10 4 spores mL −1 . One mL of spore solution was uniformly sprayed on each plant by using an airbrush compressor (Badger Airbrush model 150TM). Inoculated plants were placed for 22-24 h in a growth chamber in the dark (relative humidity ≥ 90%; 27 ± 5 • C), and further incubated in a greenhouse for disease development. Disease rating was performed 6 days after inoculation by counting the number of sporulating lesions on the youngest unfolded leaves. Photos of representative disease symptoms were taken after disease evaluation. For each treatment, three replicates of 7 plants (potting soils) or 5 plants (ASSs) each were used.

Role of Cyclic Lipopeptides in ISR Triggered by B. velezensis against P. oryzae in Potting and Acid Sulfate Soil Conditions
Based on the results of previous experiments, the most effective Bacillus isolate, namely, B. velezensis RHF4.1-25, was selected for further use and its efficacy was compared with B. velezensis GA1. In these assays cyclic lipopeptide mutants of B. velezensis GA1 were also included. Single mutants (∆srfaA, ∆ituA and ∆fenA) were tested in both potting and ASSs. Complementation of ∆ituA and ∆fenA isolates was carried out in potting soil by combining each isolate at half dose. Based on consistent results obtained in both acid sulfate soils and potting soils, and because of the limited availability of the acid sulfate soil, assays with double mutants (∆srfaA-ituA, ∆srfaA-fenA and ∆fenA-ituA) were only performed in potting soils. Bacterial inoculum preparation, experimental set up, disease infection, and evaluation were carried out in a similar manner as were performed in the assay above. For all experiments, treatments were performed in three replications, each comprising of 7 plants (potting soil) or 5 plants (ASSs).

Root Colonization Assay
Root colonization was evaluated as described previously [25,38]. At the time of disease rating, five rice roots per treatment were randomly chosen and gently washed in water to remove soil. The roots were weighed after air-drying for 2 min. Subsequently, roots were ground in 10 mL sterile saline (0.85% sodium chloride, w/v) and sterile sand by using a mortar and pestle. Following this, 100 µL of serially diluted suspensions were plated on LB agar and incubated for 24 h at 28 • C before counting Bacillus colonies based on their morphological characteristics. The data were log10 transformed prior to statistical analysis.

In Vitro Antagonism against P. oryzae Using Cell-Free Supernatants of B. velezensis Strains and GA1 CLiP Mutants
To investigate the effect of cell-free supernatants on the growth of P. oryzae mycelium, in vitro tests were carried out using plastic slides with five replications of each treatment. Sterile microscopic plastic slides covered with a thin, flat layer of water agar (Difco Bacto agar; BD Diagnostics, Le Pont-deClaix, France) were placed in a plastic Petri dish containing sterile, moist absorbent paper [39]. Subsequently, an agar plug (diameter = 8 mm) taken from an actively growing Complete Media (CM) plate of P. oryzae VT5M1 was inoculated at the center of each plastic slide. Two droplets (15 µL each) of a cell-free supernatant of the isolates were placed on two sides of the plastic slide (about 2 cm from the P. oryzae plug) while for the control treatment, two droplets of LB broth were used on both sides of the plug. All plates were incubated at 28 • C for five days. Microscopic slides were assessed by using an Olympus BX51 Microscope. Additionally, the diameter of mycelial growth was determined, and converted to the percentage growth inhibition based on the following formula: (Growth diameter of untreated control − Growth diameter of treated control) × 100 Growth diameter of untreated control 2.8. Influence of Cell-Free Culture Filtrates on P. oryzae Spore Germination and Appressoria Formation For this experiment, filter sterilized supernatants were derived from strains RHF4.1-25, GA1 and its CLiP mutants (Table 1). Spore suspensions were diluted to obtain a final concentration of 5 × 10 4 spores mL −1 and mixed with 100 µL of Bacillus cell-free supernatants to get a final concentration of 50% (v/v). The same amount of LB broth was used for the control treatment. Subsequently, a plastic slide (Fisher Scientific, Merelbeke, Belgium) containing fifty µL of the mixture was incubated in the dark at 28 • C. After a 4 h incubation period, spore germination was recorded by counting the number of germination tubes. Eight hours post incubation (hpi), fifty randomly selected spores were evaluated for appressoria formation. An Olympus BX51 Microscope was used to observe spore germination and appressoria formation. The assay was conducted twice.

Direct Effect of Cell-Free Supernatants to Reduce Rice Blast Symptoms Caused by P. oryzae Spores
To obtain 4 week old rice plants for disease infection, rice seedlings were planted in sterile potting soils (600 g per tray) and maintained under controlled conditions. Spore suspensions of P. oryzae VT5M1 were prepared as previously described [12]. Spores obtained from 8-day old P. oryzae VT5M1 cultures were added into 0.5% (w/v) gelatin to obtain a concentration of 5 × 10 4 spores mL −1 . One mL of cell-free supernatants was mixed with 1 mL of spore suspension to secure a final concentration of 50% (v/v). A compressor-powered air brush gun was used to spray the mixture onto rice plants (1 mL of the mixture per plant) and the trays were frequently rotated during this spraying process. The healthy treatment was sprayed with a mixture of 0.5% gelatin suspension and the same amount of sterile water, whereas the diseased treatment received a mixture of 0.5% gelatin and 5 × 10 4 spores mL −1 . Pathogen inoculation and disease assessment were performed as depicted above. Treatments were performed in three replications, each comprising of 7 plants.

Statistical Data Analysis
The data of all experiments were statistically analyzed using the software package SPSS 25.0. To compare mean values among treatments, univariate ANOVA followed by Duncan's post hoc tests were used and results had statistically significant differences when p < 0.05.

Potential of Indigenous Bacillus Strains to Control Rice Blast by ISR in Potting Soil and ASS Conditions
The four indigenous Bacillus strains used in this study were obtained from ASSs in Vietnam. Full details about their isolation, taxonomy and metabolic profile will be published elsewhere. Initially, they were tested for their ability to trigger ISR against rice blast disease in potting soil conditions. Three representative isolates belonging to B. velezensis (RHF4.1-25) and B. altitudinis (RHF4.1-26 and RHF3.1-20) could trigger ISR against rice blast caused by P. oryzae VT5M1, while B. marisflavi RHF2.1-7 was not effective (Figure 1). The most effective strain in triggering ISR was B. velezensis RHF4.1-25. The two isolates belonging to B. altitudinis were significantly less effective than B. velezensis RHF4.1-25. All tested Bacillus isolates effectively colonized the rice roots with densities ranging from 7.70 to 8.41 log CFU g −1 of fresh root ( Table 2).

Statistical Data Analysis
The data of all experiments were statistically analyzed using the software package SPSS 25.0. To compare mean values among treatments, univariate ANOVA followed by Duncan's post hoc tests were used and results had statistically significant differences when p < 0.05.

Potential of Indigenous Bacillus Strains to Control Rice Blast by ISR in Potting Soil and ASS Conditions
The four indigenous Bacillus strains used in this study were obtained from ASSs in Vietnam. Full details about their isolation, taxonomy and metabolic profile will be published elsewhere. Initially, they were tested for their ability to trigger ISR against rice blast disease in potting soil conditions. Three representative isolates belonging to B. velezensis (RHF4.  and B. altitudinis (RHF4.1-26 and RHF3.1-20) could trigger ISR against rice blast caused by P. oryzae VT5M1, while B. marisflavi RHF2.1-7 was not effective (Figure 1). The most effective strain in triggering ISR was B. velezensis RHF4.1-25. The two isolates belonging to B. altitudinis were significantly less effective than B. velezensis RHF4.1-25. All tested Bacillus isolates effectively colonized the rice roots with densities ranging from 7.70 to 8.41 log CFU g −1 of fresh root ( Table 2).    Figure 2). In contrast, B. marisflavi RHF2.1-7 could not trigger ISR against rice blast. Furthermore, all tested isolates colonized rice roots ranging from 6.56 until 7.54 log CFU g −1 fresh roots (Table 3).   Figure 2). In contrast, B. marisflavi RHF2.1-7 could not trigger ISR against rice blast. Furthermore, all tested isolates colonized rice roots ranging from 6.56 until 7.54 log CFU g −1 fresh roots (Table 3).

CLiP Production in B. altitudinis, B. velezensis and Mutants
UPLC-MS analysis was performed to detect and quantify CLiPs produced by B. velezensis RHF4.1-25 and its CLiP profile was compared with B. velezensis GA1 ( Figure 3, Table A2) (see Appendix A). B. velezensis RHF4.1-25 produces surfactins, fengycins and iturins, similar to the CLiPs produced by B. velezensis GA1. The CLiP profile of the various GA1 mutants including single and double CLiP mutants is also depicted in Figure 3 and Table A2. Both B. altitudinis strains RHF4.1-26 and RHF3.1-20 produce pumilacidins, as shown in Figure 3B.  Table 3. Population of Bacillus isolates applied in ISR assay on rice roots grown in acid sulfate soil ( Figure 2).

Treatment
Population Density (log CFU g −

CLiP Production in B. altitudinis, B. velezensis and Mutants
UPLC-MS analysis was performed to detect and quantify CLiPs produced by B. velezensis RHF4.1-25 and its CLiP profile was compared with B. velezensis GA1 (Figure 3, Table A2) (see Appendix A). B. velezensis RHF4.1-25 produces surfactins, fengycins and iturins, similar to the CLiPs produced by B. velezensis GA1. The CLiP profile of the various GA1 mutants including single and double CLiP mutants is also depicted in Figure 3 and Table A2. Both B. altitudinis strains RHF4.1-26 and RHF3.1-20 produce pumilacidins, as shown in Figure 3B.

Role of Cyclic Lipopeptides Produced by B. velezensis in ISR against P. oryzae VT5M1 in Potting Soil
Since B. velezensis RHF4.1-25, an indigenous isolate obtained from rice roots in Vietnam, was the most effective in triggering ISR in previous experiments, its effect was compared with the closely related isolate B. velezensis GA1, a well-studied biocontrol agent that produces the CLiPs surfactin, iturin and fengycin. To study the role of CLiPs in ISR, in a first experiment single mutants of GA1 impaired in surfactin, iturin or fengycin were included in the assay. Figure 4 shows that B. velezensis RHF4.1-25, B. velezensis GA1 wild type (GA1wt) and its surfactin mutant (∆srfaA) could trigger the ISR against rice blast caused by P. oryzae VT5M1. These treatments could significantly protect rice plants in comparison with the diseased control (DC), corresponding to approximately 30%, 54% and 46% relative infection, respectively (Figures S1 and S2). In contrast, iturin (∆ituA) and fengycin (∆fenA), mutants of B. velezensis GA1, lost the ability to trigger ISR, showing that both fengycin and iturin are needed to trigger ISR against P. oryzae VT5M1. The ISR capacity could be partially restored when both mutants were applied together (∆ituA + ∆fenA). It should be noted that in this experiment, B. velezensis RHF4.1-25 was more effective than B. velezensis GA1 (Figure 4).

Role of Cyclic Lipopeptides Produced by B. velezensis in ISR against P. oryzae VT5M1 in Potting Soil
Since B. velezensis RHF4.1-25, an indigenous isolate obtained from rice roots in Vietnam, was the most effective in triggering ISR in previous experiments, its effect was compared with the closely related isolate B. velezensis GA1, a well-studied biocontrol agent that produces the CLiPs surfactin, iturin and fengycin. To study the role of CLiPs in ISR, in a first experiment single mutants of GA1 impaired in surfactin, iturin or fengycin were included in the assay. Figure 4 shows that B. velezensis RHF4.1-25, B. velezensis GA1 wild type (GA1wt) and its surfactin mutant (∆srfaA) could trigger the ISR against rice blast caused by P. oryzae VT5M1. These treatments could significantly protect rice plants in comparison with the diseased control (DC), corresponding to approximately 30%, 54% and 46% relative infection, respectively (Figures S1 and S2). In contrast, iturin (∆ituA) and fengycin (∆fenA), mutants of B. velezensis GA1, lost the ability to trigger ISR, showing that both fengycin and iturin are needed to trigger ISR against P. oryzae VT5M1. The ISR capacity could be partially restored when both mutants were applied together (∆ituA + ∆fenA). It should be noted that in this experiment, B. velezensis RHF4.1-25 was more effective than B. velezensis GA1 (Figure 4). In this experiment the root colonization of the tested Bacillus was in the range of 10 6 CFU g −1 fresh root (Table 4). Root colonization of the CLiP mutants was not impaired in comparison with the wild type strain GA1. Table 4. Population of Bacillus isolates applied in ISR assay on rice roots grown in potting soils (Figure 4).

Treatment
CLPs In a second experiment, double mutants of B. velezensis that are impaired in the production of two of the three CLiPs were also included ( Figure 5).
All double mutants lost their ability to trigger ISR against rice blast, while the wild type strain B. velezensis GA1 and the surfactin mutant ∆srfaA were equally effective. In this experiment, B. velezensis RHF4.1-25 was significantly more effective than B. velezensis GA1. Root colonization of all isolates was in the range of 10 6 CFU g −1 of fresh root (Table 5). Table 5. Population of Bacillus isolates applied in ISR assay on rice roots grown in potting soils ( Figure 5).

Treatment
CLPs Produced Population Density (in log CFU g −

Role of Cyclic Lipopeptides in ISR of B. velezensis against P. oryzae VT5M1 in Acid Sulfate Soil
With regard to ISR against rice blast under ASS conditions, Bacillus strains including RHF4.1-25, GA1wt and only GA1 single mutants were chosen to investigate the capacity of these strains to protect the rice plants against P. oryzae VT5M1 due to the limited availability of Vietnamese ASSs to set up experiments and the consistent results above in potting soils. The results show that the ISR capacity of those isolates which could effectively protect rice plants in potting soils, worked consistently under ASS conditions. Specifically, B. velezensis RHF4.1-25, B. velezensis GA1wt and ∆srfaA isolates could significantly protect the rice plants against P. oryzae compared to the diseased control. Among these strains, RHF4.1-25 was again most effective, but not significantly different from the ∆srfaA mutant ( Figure 6). In contrast, mutants unable to produce iturin or fengycin could not significantly trigger ISR in rice blast and reduced disease severity only by 5% ( Figure S3). Furthermore, all tested isolates colonized rice roots in the range of 10 6 CFU g −1 fresh roots (Table 6).
With regard to ISR against rice blast under ASS conditions, Bacillus strains including RHF4.1-25, GA1wt and only GA1 single mutants were chosen to investigate the capacity of these strains to protect the rice plants against P. oryzae VT5M1 due to the limited availability of Vietnamese ASSs to set up experiments and the consistent results above in potting soils. The results show that the ISR capacity of those isolates which could effectively protect rice plants in potting soils, worked consistently under ASS conditions. Specifically, B. velezensis RHF4.1-25, B. velezensis GA1wt and ΔsrfaA isolates could significantly protect the rice plants against P. oryzae compared to the diseased control. Among these strains, RHF4.1-25 was again most effective, but not significantly different from the ΔsrfaA mutant ( Figure 6). In contrast, mutants unable to produce iturin or fengycin could not significantly trigger ISR in rice blast and reduced disease severity only by 5% ( Figure  S3). Furthermore, all tested isolates colonized rice roots in the range of 10 6 CFU g −1 fresh roots (Table 6).

In Vitro Antagonism of P. oryzae Using Cell-Free Supernatants of B. velezensis Wild Type Strains and CLiP Mutants
Microscopic assays were conducted to investigate direct effects of CLiP containing supernatants on the mycelial growth of P. oryzae VT5M1. Figure 7 shows that application of cell-free supernatants obtained from the wild type B. velezensis strains RHF4.1-25 and GA1 strongly suppressed the growth of P. oryzae VT5M1 with more than 94%. More importantly, abnormal hyphal fragments were also formed by those treatments (Figure 7A), as compared with that of the LB control. Supernatants containing two CLiPs (obtained from ∆srfaA, ∆ituA and ∆fenA) also strongly inhibited the growth of P. oryzae VT5M1 (from 63% to 80%), supernatants that contained both surfactin and iturin gave the best results (about 80% inhibition). Additionally, supernatants that only contained fengycin or iturin gave an intermediate inhibition from 41% to 52%. No inhibition zone was formed between fungus and supernatant that only contained surfactin ( Figure 7B).

In Vitro Antagonism of P. oryzae Using Cell-Free Supernatants of B. velezensis wild Type Strains and CLiP Mutants
Microscopic assays were conducted to investigate direct effects of CLiP containing supernatants on the mycelial growth of P. oryzae VT5M1. Figure 7 shows that application of cell-free supernatants obtained from the wild type B. velezensis strains RHF4.1-25 and GA1 strongly suppressed the growth of P. oryzae VT5M1 with more than 94%. More importantly, abnormal hyphal fragments were also formed by those treatments (Figure 7A), as compared with that of the LB control. Supernatants containing two CLiPs (obtained from ∆srfaA, ∆ituA and ∆fenA) also strongly inhibited the growth of P. oryzae VT5M1 (from 63% to 80%), supernatants that contained both surfactin and iturin gave the best results (about 80% inhibition). Additionally, supernatants that only contained fengycin or iturin gave an intermediate inhibition from 41% to 52%. No inhibition zone was formed between fungus and supernatant that only contained surfactin ( Figure 7B).    Each treatment was repeated five times and the experiment was done twice. Univariate ANOVA followed by Duncan's post hoc tests were used and different letters among these treatments indicate statistically significant differences (p < 0.05).

Influence of Cell-Free Culture Filtrates on P. oryzae Spore Germination and Appressoria Formation
This experiment was performed to study the role of CLiPs on P oryzae VT5M1 spore germination and appressoria formation. Following incubation in favorable environmental conditions for fungal growth, the percentage of conidial germination and appressoria formation in the LB control were 97% and 100%, respectively (Table 7). Interestingly, all cell-free culture filtrates obtained from wild type and mutant B. velezensis strains effectively inhibited conidial germination. More so, appressorium formation was not recorded since there was no spore germination in inoculated treatments. In addition to this, the CLiPs-containing supernatants damaged the spores and caused malformation of germ tubes (Figure 8). Table 7. Effects of cell-free culture supernatant obtained from B. velezensis RHF4.1-25, B. velezensis GA1 or its CLiP mutants on spore germination of P. oryzae VT5M1 (Figure 8).
∆srfaA-ituA: double mutant impaired in surfactin and iturin production; ∆srfaA-fenA: double mutant impaired in surfactin and fengycin production; ∆fenA-ituA: double mutant impaired in iturin and fengycin production. Each treatment was repeated five times and the experiment was done twice. Univariate ANOVA followed by Duncan's post hoc tests were used and different letters among these treatments indicate statistically significant differences (p < 0.05).

Influence of Cell-Free Culture Filtrates on P. oryzae Spore Germination and Appressoria Formation
This experiment was performed to study the role of CLiPs on P oryzae VT5M1 spore germination and appressoria formation. Following incubation in favorable environmental conditions for fungal growth, the percentage of conidial germination and appressoria formation in the LB control were 97% and 100%, respectively (Table 7). Interestingly, all cellfree culture filtrates obtained from wild type and mutant B. velezensis strains effectively inhibited conidial germination. More so, appressorium formation was not recorded since there was no spore germination in inoculated treatments. In addition to this, the CLiPscontaining supernatants damaged the spores and caused malformation of germ tubes (Figure 8). Table 7. Effects of cell-free culture supernatant obtained from B. velezensis RHF4.1-25, B. velezensis GA1 or its CLiP mutants on spore germination of P. oryzae VT5M1 (Figure 8).

Treatment
CLPs

Direct Effect of Cell-Free Supernatants to Reduce Rice Blast Symptoms Caused by P. oryzae Spores
Since significantly effective inhibition of CLiPs-producing Bacillus isolates on mycelial growth as well as on spore germination of P. oryzae VT5M1 was shown in experiments above, this assay was carried out to evaluate the capacity of the CLiPs in disease severity reduction by in planta direct antagonism. Based on the results of this plant experiment (Figure 9), application of all CLiPs-producing Bacillus isolates could significantly reduce lesion numbers compared to that of the untreated control (DC, disease control).

Direct Effect of Cell-Free Supernatants to Reduce Rice Blast Symptoms Caused by P. oryzae Spores
Since significantly effective inhibition of CLiPs-producing Bacillus isolates on mycelial growth as well as on spore germination of P. oryzae VT5M1 was shown in experiments above, this assay was carried out to evaluate the capacity of the CLiPs in disease severity reduction by in planta direct antagonism. Based on the results of this plant experiment (Figure 9), application of all CLiPs-producing Bacillus isolates could significantly reduce lesion numbers compared to that of the untreated control (DC, disease control). Microorganisms 2021, 9,   Furthermore, there were remarkable differences in disease symptom reduction between these treatments caused by spraying a 50% (v/v) concentration of mixed cell-free supernatants and spores. Interestingly, in comparison with B. velezensis GA1wt and its mutants, cell-free culture filtrates from B. velezensis RHF4.1-25 were more effective in disease suppression ( Figure S4). Among GA1wt strain and its mutants, the supernatants of GA1wt, ∆srfaA and ∆fenA provided a significant reduction in lesion numbers, as well Furthermore, there were remarkable differences in disease symptom reduction between these treatments caused by spraying a 50% (v/v) concentration of mixed cell-free supernatants and spores. Interestingly, in comparison with B. velezensis GA1wt and its mutants, cell-free culture filtrates from B. velezensis RHF4.1-25 were more effective in disease suppression ( Figure S4). Among GA1wt strain and its mutants, the supernatants of GA1wt, ∆srfaA and ∆fenA provided a significant reduction in lesion numbers, as well as relative infection, compared to that of the remaining GA1 mutants. Along with supernatantproducing ∆ituA, double mutants, namely, ∆srfaA-ituA, ∆srfaA-fenA and ∆fenA-ituA were less effective than other tested GA1 strains, but could decrease disease symptoms by 20% to 35% compared to the diseased control.

Discussion
In this study, we screened the capacity of four representative Bacillus isolates from ASSs in Vietnam to elicit ISR against rice blast disease and selected potential biocontrol agents for further studies. B. marisflavi could not successfully induce resistance to rice blast, in contrast, B. altitudinis and B. velezensis strains could effectively protect the rice plants against blast disease by inducing resistance upon root inoculation. These strains consistently triggered resistance to the blast disease in rice grown in both potting and acid sulfate soil conditions. B. velezensis RHF4.1-25 was the most successful in triggering ISR against blast disease. Our study is the first report on the capacity of indigenous Bacillus isolates to trigger ISR against blast disease on rice grown in ASSs.
The two used B. altitudinis strains could protect rice against blast disease by triggering ISR. B. altitudinis has been implicated in biocontrol in rice against the sheath blight pathogen Rhizoctonia solani by ISR [40] and against the bacterial blight pathogen Xanthomonas oryzae pv. oryzae via direct effects [41]. Both strains produce the pumilacidin-type of CLiPs. Pumilacidins are members of the surfactin family. The compounds differ from surfactin with a leucine in position 4 instead of a valine and an isoleucine or valine at position 7 instead of a leucine [42]. Pumilacidins are known to be produced by B. pumilus [43,44] and by B. safensis [45], two species closely related to B. altitudinis. To our knowledge, there are no studies showing that B. altitudinis strains produce pumilacidins or that they are active against rice blast. In this context, it should be noticed we could not demonstrate a role for surfactin produced by B. velezensis GA1 in triggering ISR against P. oryzae. However, variations in molecular structure of the peptide part can impact the physicochemical properties of a CLiP. The presence of a Leu4 in pumilacidin instead of a Val4, as in surfactin, appears to increase the critical micellar concentration value [42]. The final surface tension value of pumilacidin is also higher when compared to surfactin [45], so it cannot be excluded that this has an effect on ISR in rice. It remains to be investigated whether pumilacidins do play a role in the observed ISR against P. oryzae.
The ISR determinants of B. velezensis were deciphered in this study. Our studies have revealed that upon root inoculation, surfactin-, fengycin-and iturin-type CLiPs-producing RHF4.1-25 and B. velezensis GA1 could effectively protect rice plants against blast disease by triggering ISR. Furthermore, to shed more light on the ISR capacity of these CLiPs, single (∆srfaA, ∆ituA and ∆fenA) and double mutants (∆srfaA-ituA, ∆srfaA-fenA and ∆fenA-ituA) of GA1 were tested in this study to investigate their ISR capacity against P. oryzae VT5M1 in rice. Both fengycin and iturin are needed to trigger ISR against rice blast disease, since mutants impaired in fengycin or iturin or in both fengycin and iturin production could no longer protect the plants against rice blast disease. Additionally, mutant strains that only produce a single CLiP could not successfully protect plant against P. oryzae VT5M1 by triggering ISR. These data suggest that fengycin and iturin act synergistically to cause the ISR response, while surfactin appears not to play a role in ISR. This is in contrast with a previous study in which surfactin triggered ISR against P. oryzae in perennial ryegrass [46], indicating that the monocots rice and ryegrass do not react to lipopeptides in a similar way. Our findings are in agreement with a previous study in which surfactin and fengycin could elicit resistance against sheath blight caused by Rhizoctonia solani in rice, but were not effective against P. oryzae [31]. Another recent study showed that B. subtilis strain BJ-1, a strain isolated from a contaminated P. oryzae culture plate, could trigger ISR against rice blast on rice by seed inoculation, however, the study did not determine the effect of individual lipopeptides [15]. In many previous studies, the role of surfactin, iturin, or surfactin and fengycin together in triggering ISR in various pathosystems could be demonstrated. It has been reported that surfactin and fengycin produced by B. subtilis strain S499 could protect bean plants against Botrytis cinerea by triggering ISR [30]. Furthermore, iturin family CLiPs such as mycosubtilin and bacillomycin have also consistently triggered ISR activity against phytopathogens in various host plants [47]. These results clearly show that the capacity of CLiPs to induce resistance is pathogen and plant dependent. In addition to Bacillus, CLiP-producing Pseudomonas strains have also successfully been used to control rice blast by ISR. The capacity of the ISR-inducing Pseudomonas strains COR10, COW10 and COR5 to protect rice plants against P. oryzae was demonstrated to be due to the production of the CLiPs lokisin, WLIP and entolysin, respectively [25].
Beside indirect antagonism by inducing resistance, direct effects of cell-free supernatants produced by B. velezensis on mycelial growth, spore germination and appressorium formation of P. oryzae were also investigated. Our in vitro tests suggest that iturin and fengycin are effective in suppressing both the mycelial growth and spore germination of P. oryzae, whereas surfactin only affects spore germination. It has been shown before that surfactin produced by B. licheniformis BC98 could suppress the spore germination of P. oryzae B157 at a concentration of 1 µg mL −1 [48]. Iturin appears to be the most effective at inhibiting mycelial growth since the mutant impaired in fengycin production still showed a high relative inhibition. Surfactin-, fengycin-and iturin-producing Bacillus spp. are well-known for their direct antagonism against diverse phytopathogens in different host plants [33,47,49]. Surfactins are powerful biosurfactants that interact with lipid bilayers, and disrupt and solubilize lipid bilayers at high concentrations [33]. Fengycins and iturins directly affect fungal cell membranes and finally cause cell death [33,43,50]. Our study suggests that the presence of all three CLiPs types resulted in the best antagonistic effects on the growth of P. oryzae VT5M1. Interestingly, supernatants containing these CLiPs also completely suppressed spore germination of the fungus P. oryzae VT5M1. This result is consistent with earlier reports about the influence of surfactin-and fengycin-type CLiPs on conidial germination and appressorium formation in P. oryzae [15]. Besides, we tested the mixture of bacterial cell-free supernatants and the spore solutions as a spray on rice plants. In this case the supernatant that only contains surfactin could also significantly reduce disease severity compared to the diseased control treatment, probably because the compound could inhibit spore germination. Moreover, our results also show that the combination of surfactin-, fengycin-, iturin-type CLiPs resulted in the best protective effect on rice plants.
Interestingly, in comparison with B. velezensis GA1wt, B. velezensis RHF4.1-25 obtained from ASS could protect rice plants more effectively in both direct and indirect antagonism against blast disease caused by P. oryzae VT5M1. This is not due to a possibly better adaptation to ASS, since the difference in effectiveness was also observed in potting soil and in direct antagonism. Both strains also showed a very comparable root colonization in all plant assays. Differences in effectiveness may be due to differences in the regulation of CLiP production or in CLiP ratio. We are currently performing an in-depth genome analysis of both strains to find out why B. velezensis strain RHF4.1-25 performs better than B. velezensis GA1 in P. oryzae control.

Conclusions
Our study highlights that indigenous B. altitudinis and B. velezensis strains isolated from rice rhizosphere in ASSs in Vietnam could trigger ISR against rice blast disease. Furthermore, this study has elucidated the role of surfactin, fengycin, and iturin in controlling rice blast disease by ISR and direct antagonism. Fengycin and iturin are both needed to elicit ISR against blast disease, suggesting a synergistic interaction. Apart from ISR, our results also highlight the important role of surfactin, fengycin and iturin in the direct inhibition of P. oryzae. Both fengycin and iturin inhibit spore germination and mycelial growth, while surfactin only inhibits spore germination. This work also demonstrates that B. velezensis RHF4.1-25 isolated from rice rhizosphere in Vietnamese ASSs is more effective than B. velezensis GA1 to control rice blast by both ISR and direct antagonism, although both strains produce the same CLiPs. It is worthwhile to further explore the potential of B. velezensis strain RHF4.1-25 as biocontrol agent against rice blast in field conditions in Vietnam by applying the strain via seed application, root inoculation and/or spraying on rice leaves.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/microorganisms9071441/s1, Figure S1: potential of B. velezensis isolates RHF4.1-25 and GA1 and GA1 mutants impaired in CLiP production to induce systemic resistance against P. oryzae in rice (variety CO39) grown in potting soil. Figure S2: potential of B. velezensis isolates RHF4.1-25 and GA1, and GA1 mutants impaired in CLiP production to induce systemic resistance against P. oryzae in rice (variety CO39) grown in potting soil. Figure S3: potential of B. velezensis isolates RHF4.1-25 and GA1, and GA1 mutants impaired in CLiP production to induce systemic resistance against P. oryzae in rice (variety Jasmine 85) grown in acid sulfate soil. Figure S4: direct antagonistic activity of cell-free supernatants obtained from CLiPs-producing B. velezensis RHF4.1-25, B. velezensis GA1 wild type and its mutants against P. oryzae VT5M1.