Biological Control of Fusarium oxysporum f. sp. cubense Tropical Race 4 Using Natively Isolated Bacillus spp. YN0904 and YN1419

Fusarium wilt of banana (FWB) is the main threatening factor for banana production worldwide. To explore bacterial biocontrol resources for FWB, the antagonistic effective strains were isolated from banana-producing areas in Yunnan Province, China. Two isolates (YN0904 and YN1419) displaying strong antagonism against Tropical Race 4 (TR4) were identified from a total of 813 strains of endophytic bacteria. TR4 inhibition rates of YN0904 and YN1419 were 79.6% and 81.3%, respectively. By looking at morphological, molecular, physiological and biochemical characteristics, YN0904 was identified as Bacillus amyloliquefaciens, while YN1419 was identified as B. subtillis. The control effects of YN0904 and YN1419 on TR4 in greenhouse experiments were 82.6% and 85.6%, respectively. Furthermore, YN0904 obviously promoted the growth of banana plantlets. In addition, biocontrol marker genes related to the biosynthesis of antibiotics synthesized and auxin key synthetase genes could be detected in YN0904. Surprisingly, the marker gene sboA could be exclusively detected in YN1419, while other marker genes were all absent. Molecular characterization results could provide a theoretical basis for expounding the biocontrol mechanisms of these two strains. We concluded that natively antagonistic strains derived from local banana plantations could provide new biological control resources for FWB.


Introduction
Banana (Musa spp.) is a globally important (sub) tropical economic and food crop, grown in more than 130 countries [1,2]. China has large banana-production areas, mainly in Guangxi, Hainan, Guangdong, Fujian, Taiwan and Yunnan provinces [3]. Yunnan Province is one of the original centers of the banana and has a long history of banana cultivation. Banana production supports a key income for ethnic groups in remote areas and offers a route to alleviate poverty, playing an important economic role in regional rural society.
Fusarium wilt of banana (FWB) is a globally devastating disease caused by Fusarium oxysporum f. sp. cubense; Tropical Race 4 (TR4) is the most virulent race [4]. Being a typical soil-borne disease, TR4 is very difficult to control [5,6]. Resistant breeding is usually the most effective method; however, it is difficult to deploy conventional crossbreeding methods due to the main commercial cultivars are triploid and seedlessness [7,8]. At present, there are no fully TR4-resistant cultivars available. Some partially resistant cultivars underperform; these are more vulnerable to local conditions and possess unstable Since July 2009, healthy banana pseudostems and symptomless banana pseudostems (already naturally infected with both FWB and bacterial soft rot (Erwinia chrysanthemi)) were collected from different banana planting orchards in Yunnan Province. Isolation method of endophytic bacteria in banana pseudostems was as follows: (i). Washing the collected pseudostems with sterile water, then sterilizing with 75% ethanol by spraying after drying naturally; (ii). The 2-3 cm middle tissue of pseudostems was extracted with a sterile blade, soaked in 75% ethanol for 30 s-1 min, then transferred in 0.1% mercuric chloride for 30 s-1 min, rinsed 2-3 times with sterile water, and ground into sap with a sterile pestle and mortar; (iii). Streaking the sap on the NA medium and culturing it at 37 • C for 48 h. Then, single colonies with different morphologies were selected for purification. The isolated strains were stored at −80 • C with 50% glycerin.

Screening of Antagonistic Endophytic Bacteria against TR4
Primary screening was carried out after isolation. TR4 strain 15-1 was cultured at 28 • C for 7 days on PDA, and a 5 mm diameter slice was transferred to the center of a new PDA medium using a sterilized puncher, and then the isolated bacteria were dipped with an inoculation loop, which was 25 mm on the side of plate. After culturing at 28 • C for 7 days, the results were observed and recorded. Those strains with inhibition widths of more than 3 mm were selected for secondary screening.
The secondary screening of antagonistic bacteria was carried out according to the method of dual-culture test described by Li and He et al. [21]. The TR4 was inoculated in the center and antagonistic bacteria were inoculated at four points on cross-line graticule. The conditions of dual-culture test (including temperature and culture time) were consistent with that of initial primary screening. The PDA medium only inoculated with TR4 was used as a control, and each treatment was replicated three times. After being cultured at 28 • C for 7 days, inhibition effect results were assessed. The diameter of the TR4 colony was determined by the cross method. Inhibition rate (%) = (TR4 colony diameter of control group-TR4 colony diameter of treated group)/TR4 colony diameter of control group × 100.

Effect of Antagonistic Strains on Morphology of TR4 Hyphae
TR4 was cultured with antagonistic bacteria from secondary screening for 7 days at 28 • C using the dual-culture method. The fresh hyphae of TR4 close to the antagonistic bacteria as described by Li and He et al. [21] were selected using a sterilized toothpick and transferred to a temporary microscope slide. The effects of antagonistic bacteria on the growth of TR4 hyphae were observed using a scanning electron microscope (ZEISS Sigma 300, Berlin, Germany). The new TR4 hyphae of the single cultured were used as a control.

Morphological Observation
The antagonistic strains were cultured on NA medium at 37 • C for 24 h, then the colony morphology, transparency, color, and other characteristics were observed and recorded. Morphologies of antagonistic strains were observed using a scanning electron microscope by ZEISS Sigma 300 (Zeiss, Oberkochen, Germany).

Physiological and Biochemical Characteristics
The physiological and biochemical reactions of antagonistic strains were tested according to Fang's (1998) method [27].

Molecular Identification of Antagonistic Strains
The antagonistic strains were inoculated in NA liquid medium and then cultured at 37 • C, 180 rpm for 24 h. Bacterial genomic DNA was extracted according to the Lysis Buffer for Microorganism to Direct PCR method (TaKaRa, Dalian, China). The partial sequence of 16S rRNA gene was amplified by PCR with universal primers 27F/1492R (5 -AGAGTTTGATCCTGGCTCAG-3 /5 -GGTTACCTTGTTACGACTT-3 ) [28]; the primer sequence were synthesized by Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China). The PCR products were sent to Shanghai Personal Biotechnology Co., Ltd. for sequencing. The amplification system (25 µL) comprised: 22 µL of Golden Star T6 super PCR mix (TsingKe, Beijing, China), 1 µL of template DNA (concentration: 50-500 ng), 1 µL of upstream primer, and 1 µL downstream primer. The PCR products were sequenced, and the sequencing results were compared by blast on the NCBI website (https://www. ncbi.nlm.nih.gov/). The phylogenetic tree was constructed by the adjacency method using MEGA 7.0 software (https://www.megasoftware.net/, accessed on 23 February 2021).

Effect of Temperature on Growth of Antagonistic Strains
The antagonistic strains were activated at 37 • C for 24 h, and then the strains were inoculated in LB medium by shake culturing for 16 h, at 180 rpm. Inoculation was carried out on 50 µL bacteria suspension in a 5 mL LB medium at a concentration of 1% (v/v), which was then placed, respectively, at temperatures of 25, 28, 31, 34, 40 and 43 • C, in a shaking incubator, and finally they were cultured at 180 rpm for 12 h. An ultraviolet (UV) spectrophotometer was used to measure the absorbance value at 600 nm wavelength, and the average OD 600 value was calculated. Each treatment was replicated three times.

Effect of pH Value on Growth of Antagonistic Strains
The pH values of LB medium were adjusted to the range of 4.0, 5.0, 6.0, 7.0, 8.0, 9.0 and 10.0, respectively, using 1 mol/L hydrochloric acid solution and 1 mol/L sodium hydroxide solution. The activated strains were inoculated in LB medium by culturing at 37 • C, 180 rpm for 16 h. Inoculating 50 µL bacteria suspension into a 5 mL LB medium with the different pH values at a concentration of 1% (v/v), then incubating at 37 • C, 180 rpm for 12 h. UV spectrophotometry was used to measure the absorbance value at 600 nm wavelength, and the average OD 600 value calculated. Each treatment was replicated three times.
2.6. Greenhouse Pot Experiment 2.6.1. Pot Experiment Plantlets Greenhouse pot experiment was carried out from May to October 2020. The tissue cultured plantlets of the Brazilian banana cultivar were transferred into sand substrate to acclimatize for 30 days. The plantlets were then transplanted into plastic pots (11 cm (diameter) × 12 cm (height)) filled with banana planting substrate produced by Yunnan Yuxi Leshi Technology Co., Ltd. (Yunnan China). Banana plantlets with 5-6 leaves were selected to conduct the pot experiment to assess TR4 biocontrol and growth-promoting effects.

Preparation of Fermentation Broth
The antagonistic Bacillus strains were activated and cultured at 37 • C for 24 h; single colonies were picked using an aseptic inoculation loop and inoculated in LB medium. The fermentation broth was obtained after shaking culture at 37 • C and 220 rpm for 48 h. Sterile water was used to dilute the suspension to 1 × 10 8 cfu/mL solution and 1 × 10 7 cfu/mL fermentation solution.

Preparing the TR4 Spore Suspension
After the TR4 strain 15-1 was activated and cultured at 28 • C for 7 days on PDA, the activated hyphae were placed in PDB medium and culture was shaken at 28 • C and 180 rpm for 72 h. Then, four layers of sterile gauze were used to filter the hyphae to obtain the TR4 spore suspension. Sterile water was used to dilute the suspension to 1 × 10 6 cfu/mL spore solution of TR4.

Pot Experiment
The fermentation broth (1 × 10 8 cfu/mL, labeled as "DC1") and 10-time diluted fermentation broth (1 × 10 7 cfu/mL, labeled as "DC2") were used to drench the roots of vigorous banana plants at 40 mL per plant. After 7 days, 1 × 10 6 spores/mL TR4 spore suspension (labeled as "TR4") was applied to the banana plant roots and each plant was drenched with 40 mL. The LB culture medium without inoculated strains was used as control (CK) ( Table 1). Treatments I, II and III were inoculated with TR4, while treatments IV, V and VI were not inoculated with TR4. Each treatment was with three replicates and 4 bananas plantlets in each replicate.

Investigation of TR4 Biocontrol Effects
The biocontrol effects of Bacillus strains on TR4 in pot experiment were investigated in treatments I, II, and III, which were inoculated with TR4. Forty days after inoculation with TR4, the leaf and corm symptoms of the banana plantlets were investigated. Disease severity index (DSI) and control effect were calculated according to the literature [6,29,30].
Disease severity index = ∑ (number of diseased plants at all levels × representative value of the level)/(total number of plants × representative value of the highest level) × 100.

Evaluation of Growth-Promoting Effects of Biocontrol Bacillus Strains on Banana Plantlets
The growth-promoting effects of biocontrol strains were investigated three times: on the initial day of inoculating antagonistic bacteria, then 20 and 40 days after inoculation with TR4. The plant height, pseudostem girth and leaf number of each banana plantlet in 6 treatments were measured and recorded every time.
Plant height was measured from the ground to the intersection of the two petioles at the top by using tape measure; pseudostem girth was measured from the base of pseudostem, which was 1 cm above the ground by using vernier calipers; number of leaves was recorded as the number of green leaves per plantlet.

PCR Amplification of Biocontrol and Plant Growth-Promotion Related Genes in YN0904 and YN1419
The strains preserved in glycerin (50% sterilized glycerin mixed with equal volume of bacterial suspension) at −80 • C refrigerator were inoculated on NA medium plate by the streak-cultivation method and cultured for 14-16 h at 37 • C. Lysis Buffer for Microorganism to Direct PCR (Takara, Dalian, China) kit was used to prepare the genomic DNA of the Bacillus. The activated single colony was picked up with sterilized toothpicks and then stirred in a sterile microtube contained with 50 µL Lysis Buffer and removed. After 15 min of 80 • C thermal denaturation, centrifugation was performed at low speed, and 1-5 µL of supernatant lysate was used as follow-up PCR template.
The genomic DNA of YN0904 and YN1419 was amplified by PCR using twelve pairs of primers which target Bacillus's twelve biocontrol genes, as shown in Supplementary Table S1 and a plant growth-promotion related gene ysnE (Primers: 5 -GGCTGTAACCTTTGCTATG-3 and 5 -GCTGTTCGGGTCCTCTTTAT-3 with amplicon size: 488 bp). The PCR amplification system (total 25 µL) were 12.5 µL of MegaFTM Fidelity 2X PCR MasterMix (Applied Biological Materials Inc., Nanjing, China), 1 µL of upstream primer (10 µM), 1 µL of downstream primer (10 µM), 10.5 µL of Nuclease-free H 2 O and 1 µL of template genomic DNA. The reaction components were run for 98 • C for 30 s, followed by 35 cycles of 98 • C for 10 s, 55-62 • C for 30 s and 72 • C for 30 s, and total extending at 72 • C for 2 min. No-template reaction replacing genomic DNA with Nuclease-free H 2 O was used as negative control. B. velezensis strain YN1282-2, which is positive for all the biocontrol genes used in this study, was used as a positive control [21]. The amplified products were detected by agarose gel electrophoresis (1.2 g/100 mL).

Data Analysis
The data were analyzed by SPSS 18.0, Duncan's new complex range method and independent sample t test. The average data are expressed as "means ± standard error (S.E.)".

Isolation and Screening of Antagonistic Endophytic Bacteria
Since July 2009, a total of 256 plant samples were collected from different bananaproducing areas, including Xishuangbanna, Yuxi, Honghe, Wensahng, Baoshang and Dehong of Yunnan Province. As a result, a total of 813 strains of endophytic bacteria were isolated and 382 of them were selected for antagonism screening against TR4 (Supplementary Table S2). After the dual culture for primary screening, nine antagonistic endophytic bacteria were obtained. After the dual-culture for secondary screening, two antagonistic endophytic strains YN0904 and YN1419 with the best antagonistic effect on TR4 were obtained (Supplementary Table S2), which were isolated from symptomless pseudostem samples already infected with FWB and bacterial soft rot, and the sampling stage was at banana bud formation. YN0904 was isolated from the Musa ABB banana variety group, cv Pisang Awak, originating from Mengla (101 • 18 12 E; 21 • 27 21 N; altitude 767 m), Xishuangbanna, Yunnan in August 2009. The FWB incidence in this plantation was 3-5%. YN1419 was isolated from the Brazilian banana variety, originating from Daluo (99 • 60 18 E; 21 • 36 51 N; altitude 598 m), Xishuangbanna, Yunnan, in August 2014. The FWB incidence in this plantation was 6-8%.

Effectiveness of Antagonistic Endophytic Bacteria In Vitro
The two strains with the strongest antagonist activity against TR4 by dual culture of secondary screening were labeled as YN0904 and YN1419 (Figure 1a,b and Table S2). After being subjected to dual culture with antagonistic bacteria for 7 days, the TR4 hyphal were swollen and deformed, the tops of hyphae were expanded, and the hyphae internodes between diaphragms were shorter (Figure 1d,e). The hyphae of the TR4 control (CK) were normal, smooth and uniform ( Figure 1f). The average diameter of TR4 hyphae of YN0904 against TR4 was 1.8 cm, and the inhibition rate was 79.6% ( Table 2). The average diameter of TR4 hyphae of YN1419 against TR4 was 1.7 cm, and the inhibition rate was 81.3% (Table 2).
ter being subjected to dual culture with antagonistic bacteria for 7 days, the TR4 hyphal were swollen and deformed, the tops of hyphae were expanded, and the hyphae internodes between diaphragms were shorter (Figure 1d,e). The hyphae of the TR4 control (CK) were normal, smooth and uniform ( Figure 1f). The average diameter of TR4 hyphae of YN0904 against TR4 was 1.8 cm, and the inhibition rate was 79.6% ( Table 2). The average diameter of TR4 hyphae of YN1419 against TR4 was 1.7 cm, and the inhibition rate was 81.3% (Table 2).

Morphological Characteristics of Antagonistic Strains
The colonies of endophytic bacteria YN0904 had a light yellow color, opaque and rough surface, middle convex and irregular edge on NA medium after 24 h culture at 37 • C (Figure 2a). Further, scan electron microscopy (SEM) showed that the bacteria were rod-shaped and oval at ends and the size of bacteria was 2.3-2.5 µm × 0.5-0.8 µm (Figure 2b). The colonies of endophytic bacteria YN1419 had a grayish-white color, convex, irregular edge, opaque and rough surface on NA after 24 h of culturing at 37 • C (Figure 2c). SEM showed that the bacteria were rod-shaped and oval at ends; the size of bacteria was 2-3 µm × 0.8-1 µm. (Figure 2d). rough surface, middle convex and irregular edge on NA medium after 24 h culture at 37 °C (Figure 2a). Further, scan electron microscopy (SEM) showed that the bacteria were rod-shaped and oval at ends and the size of bacteria was 2.3-2.5 μm × 0.5-0.8 μm ( Figure  2b). The colonies of endophytic bacteria YN1419 had a grayish-white color, convex, irregular edge, opaque and rough surface on NA after 24 h of culturing at 37 °C (Figure 2c). SEM showed that the bacteria were rod-shaped and oval at ends; the size of bacteria was 2-3 μm × 0.8-1 μm. (Figure 2d).

Physiological and Biochemical Characteristics of Antagonistic Strains
Physiological and biochemical tests of strains YN0904 and YN1419 are shown in Table 3. YN0904 was identified as a Gram-positive bacterium, and showed positive to the test of sucrose, α-D-Glucose, amylose hydrolysis, etc. The isolate YN1419 was also identified as a Gram-positive bacterium, and tested positive for sucrose, α-D-Glucose, amylose hydrolysis, etc. We also determined the optimal temperature and pH for YN0904 and YN1419 growth were 37 °C and 7.0, respectively ( Figures S1 and S2).

Physiological and Biochemical Characteristics of Antagonistic Strains
Physiological and biochemical tests of strains YN0904 and YN1419 are shown in Table 3. YN0904 was identified as a Gram-positive bacterium, and showed positive to the test of sucrose, α-D-Glucose, amylose hydrolysis, etc. The isolate YN1419 was also identified as a Gram-positive bacterium, and tested positive for sucrose, α-D-Glucose, amylose hydrolysis, etc. We also determined the optimal temperature and pH for YN0904 and YN1419 growth were 37 • C and 7.0, respectively ( Figures S1 and S2).

Molecular Identification of Antagonistic Bacteria
When aligning the 16S sequence of antagonistic strain YN0904 (GenBank Accession No. MW647760) in GenBank, the highest similarity sequence was with B. amyloliquefaciens. The 16S rRNA gene fragment of YN0904 was 1420 base pairs (bp). The phylogenetic tree based on 16S rRNA gene sequence results showed that strain YN0904 and B. amyloliquefaciens (MT613661, 65777, MW082822) clustered in the same large branch with a 69% bootstrap value ( Figure 3). As a conclusion, YN0904 was identified as B. amyloliquefaciens based on its molecular, morphology, physiological and biochemical characteristics.

Molecular Identification of Antagonistic Bacteria
When aligning the 16S sequence of antagonistic strain YN0904 (GenBank Accession No. MW647760) in GenBank, the highest similarity sequence was with B. amyloliquefaciens. The 16S rRNA gene fragment of YN0904 was 1420 base pairs (bp). The phylogenetic tree based on 16S rRNA gene sequence results showed that strain YN0904 and B. amyloliquefaciens (MT613661, 65777, MW082822) clustered in the same large branch with a 69% bootstrap value (Figure 3). As a conclusion, YN0904 was identified as B. amyloliquefaciens based on its molecular, morphology, physiological and biochemical characteristics. When aligning the 16S sequence of the antagonistic strain YN1419 (GenBank Accession No. MW647761) in GenBank, the highest degree of sequence similarity was with B. subtilis. The 16S rRNA gene fragment of YN1419 was 1419 base pairs (bp). The phylogenetic tree based on 16S rRNA gene sequence results showed that this strain and B. subtilis (MT538531, MT513998) clustered in the same large branch with a 99% bootstrap value ( Figure 3). As a conclusion, YN1419 was identified as B. subtilis based on its molecular, morphological, physiological and biochemical characteristics. When aligning the 16S sequence of the antagonistic strain YN1419 (GenBank Accession No. MW647761) in GenBank, the highest degree of sequence similarity was with B. subtilis. The 16S rRNA gene fragment of YN1419 was 1419 base pairs (bp). The phylogenetic tree based on 16S rRNA gene sequence results showed that this strain and B. subtilis (MT538531, MT513998) clustered in the same large branch with a 99% bootstrap value ( Figure 3). As a conclusion, YN1419 was identified as B. subtilis based on its molecular, morphological, physiological and biochemical characteristics.

YN0904 and YN1419 Had Good Biocontrol Effects on TR4 in Pot Experiment
At 40 days post inoculation (dpi), the leaves of the control group (CK + TR4) that had been inoculated only with TR4 turned yellow and the plantlets were stunted, while most leaves of the treated groups DC1 of both YN0904 and YN1419 remained healthy (Figure 4a). Additionally, after splitting the corm, we found that the most corms of treated groups DC1 and DC2 of both YN0904 and YN1419 were hardly invaded by TR4 hyphae, while TR4 caused severe damage in the control-the color of the corms was red or brownish-black (Figure 4b).
At 40 days post inoculation (dpi), the leaves of the control group (CK + TR4) that had been inoculated only with TR4 turned yellow and the plantlets were stunted, while most leaves of the treated groups DC1 of both YN0904 and YN1419 remained healthy ( Figure  4a). Additionally, after splitting the corm, we found that the most corms of treated groups DC1 and DC2 of both YN0904 and YN1419 were hardly invaded by TR4 hyphae, while TR4 caused severe damage in the control-the color of the corms was red or brownishblack (Figure 4b).  Disease investigation showed that there was a significant TR4 suppressive effect of the two antagonistic bacteria. For the disease indexes, the treated groups DC1 and DC2 of both YN0904 and YN1419 on banana corms and leaves were significantly lower than that in treated groups. Among them, there was no significant difference between YN0904 DC1 and DC2, but YN1419 DC1 was significantly lower than DC2 (Figure 4a,b and Table 4). Disease investigation showed that there was a significant TR4 suppressive effect of the two antagonistic bacteria. For the disease indexes, the treated groups DC1 and DC2 of both YN0904 and YN1419 on banana corms and leaves were significantly lower than that in treated groups. Among them, there was no significant difference between YN0904 DC1 and DC2, but YN1419 DC1 was significantly lower than DC2 (Figure 4a,b and Table 4). For the biocontrol effects, the treated groups DC1 of YN0904 on banana corms and leaves were slightly higher than DC2, respectively, but the difference between them was not significant. The treated groups DC1 of YN1419 on banana corms and leaves were significantly higher than DC2. There was no significant difference between the treated groups DC1 of YN0904 and YN1419 on banana corms and leaves, respectively (Figure 4a,b and Table 4).

Growth-Promoting Effect of Antagonistic Bacteria on Banana Plantlets
The inoculation of two biocontrol strains YN0904 and YN1419 did not affect the growth of banana plantlets, and YN0904 also had obvious growth-promoting effects on banana plants (Figure 4c).

CK + TR4
56.25 ± 3.61 a 70.83 ± 2.08 a Data are presented as means ± standard error. Data with different lowercase letters indicate a significant difference at the 0.05 level.
For the biocontrol effects, the treated groups DC1 of YN0904 on banana corms and leaves were slightly higher than DC2, respectively, but the difference between them was not significant. The treated groups DC1 of YN1419 on banana corms and leaves were significantly higher than DC2. There was no significant difference between the treated groups DC1 of YN0904 and YN1419 on banana corms and leaves, respectively (Figure 4a,b and Table 4).

Growth-Promoting Effect of Antagonistic Bacteria on Banana Plantlets
The inoculation of two biocontrol strains YN0904 and YN1419 did not affect the growth of banana plantlets, and YN0904 also had obvious growth-promoting effects on banana plants (Figure 4c).

PCR Amplification Results of Biocontrol and Plant Growth-Promotion Related Genes in YN0904 and YN1419
For twelve biocontrol related genes, seven non-ribosomal peptide synthetases (NRPS) genes including fenD, ituC, yngG, yndJ, srfAA, bamD, and dhb, which are responsible for the syntheses of seven non-ribosomal metabolites, were presented in YN0904, except for srfAA (Table 5 and Figure S3). Among the four polytide synthetases (PKS) genes,  , 6b and 7b).

PCR Amplification Results of Biocontrol and Plant Growth-Promotion Related Genes in YN0904 and YN1419
For twelve biocontrol related genes, seven non-ribosomal peptide synthetases (NRPS) genes including fenD, ituC, yngG, yndJ, srfAA, bamD, and dhb, which are responsible for the syntheses of seven non-ribosomal metabolites, were presented in YN0904, except for srfAA (Table 5 and Figure S3). Among the four polytide synthetases (PKS) genes, including dfn, bae, mln and bac, which were responsible for the syntheses of four polyketides metabolites, YN0904 contains dfn, bae and bac genes (Table 5 and Figure S3). Surprisingly, there are no NRPS and PKS genes in the YN1419 genome (Table 5 and Figure S3). Only the marker gene sboA was detected in YN1419, while it was not detected in YN0904 (Table 5 and Figure S3).
The plant growth-promotion related gene ysnE, which is responsible for the synthesis of auxin, was only detected in YN0904 (Table 5 and Figure S4).
The genomic DNA of Bacillus strains YN10904 and YN1419 was amplified respectively by PCR using thirteen pairs of published primers which target twelve biocontrol marker genes and a plant growth-promotion gene ysnE. "+" represents the positive result of PCR, "−" represents the negative result of PCR.

Discussion
The chemical components of some fungicides have a disinfecting effect on TR4 in vitro, but the control effect in soil environment is not clear, and it is unfriendly to the environment [31]. At present, chemical control measures have no significant effect on the control of FWB in the field, and biological control is the most potential and effective method [1]. Endophytic bacteria are a kind of potential biocontrol resource, which can secrete a variety of antibacterial substances and reproduce independently in plants [32,33]. In this study, further antagonistic determination of the isolated strains was conducted; the endophytic bacterial strains YN0904 and YN1419 showed the TR4 antagonistic effect in vitro ( Figure 1 and Table S2), which also showed significant suppressive effects on TR4 in vivo (Figure 4a,b and Table 4). These two strains have no deleterious effects on banana growth and YN0904 also has an obvious growth-promoting effect (Figures 4c and 5-7). YN0904 could produce a variety of antibiotics to inhibit the growth of TR4, which could alleviate the symptoms of plants. The strain YN0904 contains IAA synthesis genes (Table 5), which would synthesize IAA and promote to the normal growth of plants after being infected by TR4. Surprisingly, the marker gene sboA could only be detected in YN1419 while the other 11 marker genes are all absent ( Figure S4). Subtilosin A was proved as a broad-spectrum antibiotic against bacterial pathogens, but its anti-fungal effect has not been reported. In addition, the ysnE gene could not be detected in the YN1419 genome ( Figure S4). YN1419 may antagonize TR4 and promote plant growth through the synthesis of new substances or through new biocontrol mechanisms.
By morphological, molecular identification, physiological and biochemical characteristics, two strains, YN0904 and YN1419, were identified as B. amyloliquefaciens and B. subtillis, respectively. Bacillus is a family of bacteria with rich diversity, which has many qualified characteristics such as stress resistance, broad-spectrum antibacterial, environmental friendly, low pathogenicity to plants, and resilient adaptability properties [34]. The biocontrol Bacillus could produce abundant secondary metabolites during the growth process to inhibit the activity of pathogenic fungi or bacteria and have a promoting effect on plants growth [35]. They are considered to be ideal biocontrol agent against soil-borne diseases such as FWB [36][37][38].
At present, research on biocontrol of FWB are mainly focused on morphology assays or pot experiments; the efforts to explore the antagonistic mechanisms of biocontrol strains are limited. B. amyloliquefaciens is a kind of main bacterium used in biological control of FWB [39][40][41]. In this study, the control effect of endophytic B. amyloliquefaciens YN0904 on FWB was investigated both on leaves and corm of banana plants. The control effect of YN0904 is better than other endophytic B. amyloliquefaciens reported in previous studies. Zhu et al. [42] have reported that endophytic B. amyloliquefaciens had good control effects against Fusarium chlamydospore. Marcos et al. [43] reported an endophytic B. amyloliquefaciens played a functional role in enhancing growth and disease protection of invasive English ivy (Hedera helix). B. subtillis is a similar species to B. amyloliquefacien. It usually has a broad-spectrum antibacterial activity and growth-promoting effects. It is widely used in biological control of plant diseases in agricultural production [44,45]. There are few reports about the control effect of B. subtilis on FWB. In this study, the control effect of endophytic bacteria B. subtilis YN1419 on FWB was obvious.
Previous studies have shown that in the lipopeptide metabolites synthesized by Bacillus, iturin and fengycin had strong antagonistic activity against pathogenic fungi, and there was a synergistic effect between iturin and fengycin [46,47]. Bacillomycin D and fengycin were related to the antagonistic effect of B. amyloliquefaciens SQR9 on TR4 [48]. The crude extract of BEB17 contains bacillibactin, difficidin and bacillaene, etc., which had a good inhibitory effect on TR4 [49]. Li and He et al. detected 12 biocontrol marker genes by PCR in five Bacillus strains with good TR4 antagonistic effects and found that most of the biocontrol genes could be detected in these five strains, and they speculated that the antibacterial effect of the five Bacillus strains might be due to the presence of key biosynthesis genes of these biocontrol metabolites in their genomes [21]. In this study, most of the biocontrol marker genes were detected in the genomes of YN0904, except for srfAA, mln and sboA (Table 5 and Figure S3). We speculated that the biocontrol effect of YN0904 on TR4 may be related to the synthesis of these substances. Surprisingly, only the sboA gene was detected in the YN1419 genome, even its biocontrol effect on TR4 was slightly better than that of YN0904. However, the corresponding antibiotic of sboA is subtilosin A, which is a broad-spectrum antibiotic against bacterial pathogens, but its resistance to fungi has not been reported so far [50]. Therefore, it is likely that there are some new antagonistic substances or new antagonistic mechanisms in YN1419, but it needs further confirmation in our next step of research. In addition, ysnE, a key IAA synthase gene [25], was detected in YN0904 (Table 5 and Figure S4), indicating that the plant growth-promoting effect of YN0904 when plant was infected with TR4 may be related to the biosynthesis of auxin. However, the ysnE gene is not present in the YN1419 genome, but it has a significant growth-promoting effect on TR4 infected banana. We will sequence the whole genome of YN1419 next to further explore the TR4 biocontrol and banana growthpromoting mechanism of YN1419, and the specific synthesis genes of these biocontrol and growth-promoting substances need to be further verified by specific gene knock-out in the next step of our research.
The banana industry offers a key route to poverty alleviation and increasing local farmers' incomes in remote areas of Yunnan for ethnic groups. However, the current banana industry is seriously threatened by TR4. After years of screening antagonistic bacteria against FWB, two strongly antagonistic strains belonging to endophytic bacteria have now been identified from local banana plantations. The mechanisms for disease suppression by antagonistic strains YN0904 and YN1419, and monitoring their field-level control effects need to be further studied and validated. Furthermore, these two natively strains could be developed into biofertilizer in subsequent steps of our research because they are already adapted to local ecological conditions, which is a promising way to solve the problem and contributes to the stable development of the banana industry.

Conclusions
Two TR4 antagonistic endophytic bacteria were identified from banana plants, which were identified as B. amyloliquefaciens and B. subtilis by morphological and molecular biology and labeled as YN0904 and YN1419, respectively. The pot experiment showed that YN0904 and YN1419 had significant inhibitory effects on FWB, and these two strains had no significant deleterious plant growth effect on banana growth. The PCR results showed that several biocontrol marker genes related to the biosynthesis of antibiotics synthesized by NRPS, PKS and RPS pathways and IAA key synthetase gene could be detected in the genomes of YN0904. However, the genomes of YN1419 contain only one biocontrol gene sboA, which is a key synthase gene of subtilosin A. Molecular characterization provided a theoretical basis for expounding the biocontrol mechanisms of these two strains.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/jof7100795/s1, Figure S1: Effect of temperature on the growth rate of strains YN0904 and YN1419. Figure S2: Effect of temperature on the growth rate of strain YN0904 and YN1419. Figure S3: PCR detection results of NRPS, PKS and RPS biological control genes in YN0904 and YN1419; Figure S4: PCR detection results of plant auxin synthase gene ysnE in YN0904 and YN1419; Table S1: 12 pairs of primers of Bacillus's biocontrol genes; Table S2: Background information of isolated endophytic bacteria.