Introduction

Plant diseases are a major and long-term threat to crop yield and quality worldwide (Compant et al. 2005). Fungicides have been major method to control pathogen over the past decades. Continuous use of fungicides leads to fungicide resistance, causes environmental pollution, and brings risks to food safety and human health (Shafi et al. 2017). The use of natural antagonistic microorganisms as microbial inoculants is an ideal alternative or a supplemental way to control pathogens. Bacillus species are effective against a broad range of pathogenic microorganisms (Berg 2009). Bacillus velezensis with characteristics of fast growth and stability is widely distributed in nature and plays an increasingly important role in the fields of agriculture (Ye et al. 2018). B. velezensis as heterotypic synonyms of B. amyloliquefaciens subsp. plantarum is distinguished from B. amyloliquefaciens by secondary metabolite production, comparative genomics, and DNA-DNA relatedness calculations, and these two species are separated from B. subtilis (Dunlap et al. 2016; Fan et al. 2017; Andrić et al. 2020; Rabbee et al. 2019). More and more studies have investigated the potential of B. velezensis for curing or preventing plant diseases. For example, B. velezensis strain ZSY-1, isolated from Chinese catalpa, inhibited the growth of Alternaria solani and Botrytis cinerea by volatile organic compounds (Gao et al. 2017). B. velezensis strains 5YN8 and DSN012 controlled pepper gray mold disease by suppressing mycelium growth and spore formation (Jiang et al. 2018). B. velezensis C2, isolated from the crown tissue of tomato, exhibited significant antifungal activity against Verticillium dahliae through secondary metabolites and lytic enzymes (Dhouib et al. 2019).

With the increasing number of Bacillus species isolated, antimicrobial substances and fungal antagonistic mechanisms have been gradually explored (Lopes et al. 2018). Bacillus spp. are able to control plant diseases through diverse mechanisms, including producing antimicrobial compounds, competition with pathogens for space and nutrients, stimulation the induced systemic resistance (ISR) of plant, and promotion of plant growth (Fan et al. 2018; Shafi et al. 2017). B. velezensis harbors a high genetic capacity for synthesizing secondary metabolites, playing important roles in pathogen suppression. For example, gene clusters encoding surfactin (srf), fengycin (fen), macrolactin (pks2), bacillaene (bae), difficidin (dfn), bacilysin (bac), and bacillibactin (dhb) were present in model fungal antagonistic bacterium B. velezensis FZB42 (Fan et al. 2018). Various metabolic substances exert fungal antagonistic effects through different mechanisms. For instance, B. velezensis RC 218 showed antagonist activity against Fusarium graminearum due to direct antagonism by secondary metabolites (Palazzini et al. 2016). Lipopeptides surfactin and fengycin can act as elicitors of induced systemic resistance in plants (Chen et al. 2020). Surfactin is also essential for root colonization and influenced the ecological fitness (Ongena and Jacques 2008). Siderophore bacillibactin is involved in regulation of ferric ion (Khan et al. 2018). Nevertheless, the fungal antagonistic mechanisms still need to be further elucidated. Genome sequencing, genome annotation, and comparative genome analysis are important approaches to provide insight into the biology of fungal antagonistic strains.

In this study, Bacillus strain LJBV19 was isolated from rhizosphere soil of grapevine and evaluated against 12 phytopathogens such as Magnaporthe oryzae, Colletotrichum gloeosporioides, and Fusarium solani. In view of morphological, physicochemical, molecular analysis and genome comparison, strain LJBV19 belonged to B. velezensis. The whole genome of LJBV19 was sequenced and annotated to explore its fungal antagonistic mechanisms. LJBV19 genome was compared with three close strains B. velezensis FZB42, B. amyloliquefaciens DSM7T, and B. subtilis 168T. Genome comparison showed common and unique gene clusters related to the biosynthesis of secondary metabolites in LJBV19 genome. Overall, our data indicated that LJBV19 had the potential for protecting plant health against a broad range of pathogens.

Materials and methods

Isolation of bacteria

In October 2019, rhizosphere soil sample was collected from vineyard in Wujing town, Minhang district, Shanghai, China. Sample was dissolved in NaCl solution (0.85% w/v) and vibrated violently for 2 min (Santana et al. 2008). One milliliter of soil solution was incubated at 80 °C for 30 min, and then was diluted tenfold, 100-fold, and 1000-fold. After the solution cooled on ice, 0.1 mL liquid from each dilution was uniformly coated on Luria–Bertani (LB) agar medium (5.0 g/L yeast extract, 10.0 g/L peptone, 10.0 g/L NaCl, and 15.0 g/L agar, pH neutral), and was incubated at 37 °C for 12 h. Single clones with biofilm were randomly selected for streaking purification according to Sari et al. (2019). One clone, named as LJBV19, which obviously inhibited the hyphae growth of C. gloeosporioides, Coniothyrium diplodiella, and B. cinereal, was maintained on LB slants at 4 °C and stored with glycerol at −20 °C for further study.

Morphological, physiological, and biochemical analysis

After incubating on LB agar plate at 37 °C for 24 h, the colony characters of LJBV19 were recorded. Gram staining and spore staining were performed using Gram Stain Kit (Beijing Solarbio Science & Technology Co., Ltd) and Spore Stain Kit (Solarbio), respectively. The biochemical characteristics were identified using traditional approaches according to the Bergey’s Manual of Systematic Bacteriology (Anonymous 2001). LB liquid medium fermentation broth was centrifuged after incubated at 37 °C for 48 h, and the supernatant was used for determining defense-related enzyme activities. Cellulase, chitinase, and chitosanase activities were measured using 3,5-dinitro salicylic acid (DNS) method (Zhu et al. 2007). One unit (U) of chitinase, chitosanase, and cellulase resulted in 1 μmol of D-glucose, N-acetyl-D-glucosamine, and D-glucosamine per min, respectively.

Antifungal spectrum analysis

The antifungal spectrum of LJBV19 against plant pathogens was performed through plate assays on potato dextrose agar (PDA) according to the previous method (Wang et al. 2021). Briefly, LJBV19 was streaked horizontally in the center of the 9-cm-diameter PDA agar (boil chopped potatoes at 200 g/L for 30 min then filter with gauze and discard the residue, add 20 g/L glucose, and 20 g/L agar, pH neutral) plate. Mycelial plugs from the margin of the pathogen colony were placed on the left and right sides 2.2 cm from the center of plate. One mycelial plug was placed on each side of the center of plate, and the two plugs were removed from the same pathogen. LB medium was used as the control. The plates were incubated in the dark for 7 days for fungal growth. The inhibition activity was defined as the percentage of mycelial growth inhibition and calculated using the following formula: inhibition (%) = ((R1 − R2)/R1) × 100% (Zhang et al. 2016). R1 and R2 were the radius of the mycelium in the control and treatment, respectively.

Genome sequencing and assembly

The genome of LJBV19 was sequenced by Personalbio Technology Co., Ltd., Shanghai, China. TIANamp Bacteria DNA Kit provided by Tiangen Biotech (Beijing) Co., Ltd. was used to extract the genomic DNA of LJBV19. The qualified genomic DNA of LJBV19 was fragmented with G-tubes for Oxford Nanopore Technologies (ONT) Library. Then, DNA fragmentations were treated for damage repair and end repair, adaptor ligation, and size selection with a BluePippin system to prepare ONT libraries. ONT Library quality was detected by Qubit and sequencing was performed by ONT platform according to standard protocols. The reads of the ONT were assembled de novo using Hierarchical Genome Assembly Process (HGAP) (Chin et al. 2016). Libraries for Illumina PCR-free paired-end genome sequencing were constructed according to Illumina TruSeq DNA Sample Preparation Guide. The genomic DNA was fragmented using Covaris. DNA fragments were treated for double-end repair and sequencing adapter ligation. After quality control, the PCR-free libraries were sequenced using paired-end sequencing by Illumina NovaSeq platform. Utilizing Illumina short reads, software Pilon (Walker et al. 2014) was used to correcting the errors in ONT long-read assembly and improve the accuracy of the sequence.

Genomic feature prediction and annotation

The ORF of LJBV19 genome was predicted using GeneMarkS (Besemer et al. 2001). tRNA genes were predicted by tRNAscan-SE (Lowe and Eddy 1997), and rRNA genes were carried out by RNAmmer (Lagesen et al. 2007). Other non-coding RNA, such as small nuclear RNAs (snRNAs), was predicted by BLAST searching against the Rfam database (Kalvari et al. 2018). The functions of genes were predicted through comparisons against diverse protein databases, including Gene Ontology (GO) (Ashburner et al. 2000), Swiss-Prot (Boeckmann et al. 2003), Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa et al. 2016), the enhanced Cluster of Orthologous Groups of proteins (eggNOG) (Jensen et al. 2008), and Non-Redundant Protein Database (NR) (Li et al. 2002). CGView was used to generate the graphical view of LJBV19 genome (Chin et al. 2016). SignalP (Bendtsen et al. 2004) was used to annotate signal peptides, and TMHMM (Chen et al. 2003) was used to annotate proteins with transmembrane structure. Proteins containing the signal peptide structure without transmembrane structure were secreted proteins. The CAZy database was used to further analyze carbohydrate active enzymes (CAZymes) (Lombard et al. 2014). Additional annotations were performed by the following software: IslandViewer (Bertelli et al. 2017), hmmscan (Choo et al. 2004), and CRISPR finder (Grissa et al. 2008).

Phylogenetic analysis and comparative genomic analysis

27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-ACGGCTACCTTGTTACGACTT-3′) primers were used to amplify 16S rRNA. The PCR product was sequenced by Shanghai Sunny Biotechnology Co., Ltd. The 16S rRNA sequence of LJBV19 was deposited in GenBank (Accession No. MZ157279) and compared in the EZBiocloud (https://www.ezbiocloud.net/) database. MEGA 7.0 software was used for phylogenetic analysis using neighbor joining method. The average nucleotide identity (ANI) and digital DNA: DNA hybridization (dDDH) were analyzed by Jspecies (http://jspecies.ribohost.com/jspeciesws/) and Genome-to-Genome Distance Calculator (GGDC) (https://ggdc.dsmz.de/ggdc.php) (Richter and Rossello-Mora 2009), respectively. Four closely related Bacillus species with released complete genomes, including B. velezensis FZB42 (GenBank: NC_009725.2) (Chen et al. 2007), B. amyloliquefaciens DSM7T (GenBank: NC_014551.1) (Borriss et al. 2011), and B. subtilis 168T (GenBank: NC_000964.3) (Borriss et al. 2018), were selected for genome comparation by Mauve using the progressive alignment, and the LJBV19 genome served as the reference genome (Darling et al. 2010). R package was used to generate Venn diagram (Richter and Rossello-Mora 2009). Furthermore, the putative secondary metabolite clusters were identified using the antiSMASH v6.0 (https://antismash.secondarymetabolites.org) program with default parameters (Blin et al. 2021). Comparative analyses of gene clusters identified in B. velezensis LJBV19, B. velezensis FZB42, B. amyloliquefaciens DSM7T, and B. subtilis 168T were performed by KEGG (Kanehisa et al. 2016) and the GenBank database.

Nucleotide sequence accession numbers

The complete genome sequence of B. velezensis LJBV19 strain was deposited in the GenBank database under the accession number CP072563. This strain is available with the “China General Microbiological Culture Collection Center” (CGMCC), Beijing, China, under the accession number CGMCC No. 21804.

Results and discussion

Organism information

LJBV19 was a gram-positive, endospore-forming, rod-shaped, and aerobic bacterium. This strain grew on LB agar at 37 °C for 24 h, producing nearly round and creamy white colonies with irregular margins and dry wrinkles on the surface (Fig. S1). LJBV19 could grow in 8.5% (w/v) NaCl and over a wide pH range (4.5–10.0). LJBV19 was positive for catalase, Voges-Proskauer, nitrate reduction, gelatin liquefaction, and hydrogen sulfate test (Table S1). Besides, strain LJBV19 could utilize diverse carbon sources, including starch, citrate, xylose, sucrose, and multiple monosaccharides (Table S1). All tests indicated that the morphology, physiological, and biochemical characteristics of LJBV19 were similar to Bacillus species. Minimum information about the genome sequence (MIGS) of LJBV19 was listed in Table S2.

Taxonomic position of LJBV19

Compared to sequences of the type strains of B. velezensis, B. subtilis, and B. amyloliquefaciens, the 16S rRNA identity of LJBV19 were 100, 99.70, and 99.70%, respectively. To understand the systematic classification of LJBV19, a phylogenetic tree was constructed based on16S rRNA gene (Fig. S2). However, it was difficult to differentiate B. velezensis, B. amyloliquefaciens, and B. subtilis according to traditional phenotypic and similarity analysis of 16S rRNA (Rooney et al. 2009). With the development of sequencing technology and bioinformatics, many approaches such as ANI and silico DDH analyses have been used to differentiate and re-categorized species in Bacillus taxa (Cai et al. 2017). For example, B. velezensis FZB42 was previously grouped as B. amyloliquefaciens (Adeniji et al. 2019).

To further clarify the taxonomic position of LJBV19, ANI and DDH analyses were performed (Table S3). ANI and dDDH values between strains LJBV19 and B. velezensis NRRL B-41580 were 98.89 and 92.1%, respectively. Similarly, ANI and dDDH value were 98.17 and 85.5% compared to B. velezensis FZB42, respectively. The ANI and dDDH values between LJBV19 and B. amyloliquefaciens DSM7T were 93.39 and 56.0%, respectively. In addition, there were lower ANI and dDDH values when compared with B. subtilis 168T. According to ≥ 95% similarity in ANI and ≥ 70% homology in dDDH belonging to the same species (Chun et al. 2018), LJBV19 was affiliated with B. velezensis. Overall, results of ANI and DDH consistent with the result in the 16S rRNA tree showed that LJBV19 should be classified as B. velezensis.

Data information about the public accessibility of all material

Bacillus sp. LJBV19 genome assembly ASM1779784v1, submitted by Shanghai Jiao Tong University. April, 2021. RefSeq: GCF_017797845.1, GenBank: GCA_017797845.1.

Fungal antagonistic effect of LJBV19

There were diverse genera of microorganisms in rhizosphere, which had important effects on plant growth, pathogen defense, and resistance (Sasse et al. 2018). In the study, we have screened LJBV19 with broad spectrum antagonistic activities from rhizosphere soil to enhance plant disease resistance. Plate co-culture assay showed that LJBV19 could inhibit the mycelia growth of diverse pathogens (Fig. 1). The growth of M. oryzae was significantly suppressed with an inhibition ratio up to 75.55%. Strain LJBV19 showed good inhibition (> 30%) towards C. gloeosporioides, F. solani, Verticiltium dahlia, Exserohilum rostratum, Phytophthora capsica, and F. graminearum (Table S4). In addition, the mycelial growth of B. cinerea, Fusarium equiseti, C. diplodiella, Fusarium oxysporum, and Rhizoctonia solani were influenced by LJBV19 in varying degrees (Table S4). Importantly, the mycelia on the antagonistic plate, such as B. cinerea, were significantly enlarged, twisted, and broken under microscopic observation (Fig. S3B and C). The mycelia in the control grew normally (Fig. S3A). These results indicated that strain LJBV19 had broad spectrum antimicrobial activity to fungal phytopathogens. Meanwhile, most researches supported the fungal antagonistic potential of B. velezensis, such as B. velezensis CC09 inhibiting wheat powdery mildew, B. velezensis BAC03 as an effective antagonist of Streptomyces scabies, and B. velezensis BS87 and RK1 as bioprotection agents of strawberries against F. oxysporum (Adeniji et al. 2019).

Fig. 1
figure 1

Effect of B. velezensis LJBV19 on growth of 12 plant pathogens. A and a Magnaporthe oryzae. B and b Colletotrichum gloeosporioides. C and c Fusarium solani. D and d Verticiltium dahlia. E and e Exserohilum rostratum. F and f Phytophthora capsica. G and g Fusarium graminearum. H and h Botrytis cinerea. I and i Fusarium equiseti. J and j Coniothyrium diplodiella. K and k Fusarium oxysporum. L and l Rhizoctonia solani. Uppercase letter indicated the treatment and lowercase letter indicated the control

Full size image

Genome features of LJBV19

The complete genome of LJBV19 contained a circular 3,973,013 bp chromosome with 43.96% GC, which were within the genome size range of 3.81–4.24 Mbp and 45.9–46.8% GC content reported for this species (Mullins et al. 2020). A graphical circular map of the genome showing the genome structure and functions was presented in Fig. 2. There were 3993 open reading frames (ORFs) predicted by GeneMarkS in the genome of LJBV19. In addition, 27 rRNA genes, 86 tRNA genes, and 50 pseudogenes were contained in LJBV19 genome. Using the SignalP, and TMHMM databases, 210 (5.26%) and 1,012 (25.34%) of ORFs were divided into encoding signal peptides and transmembrane helices, respectively. One hundred and five (2.63%) proteins contained the structure of signal peptides without transmembrane helices, secreted proteins, were predicted. Besides, 248 genomics islands (GI), 13 virulence factors of pathogenic bacteria (VFDB), and 8 prophage regions were present in the genome of LJBV19. The functions of genes that were predicted using various databases showed that 3836 (96.07%), 3308 (82.84%), 2181 (54.62%), 3493 (87.48%), and 2728 (68.32%) ORFs matched in the NR, eggNOG, KEGG, SwissProt, and GO databases, respectively. In the eggNOG database, 3308 ORFs were classified into 19 COG categories, including 2.08% related to secondary metabolites biosynthesis, transport, and catabolism (Q); 5.84% to carbohydrate transport and metabolism (G); 6.69% to amino acid transport and metabolism (E); and 6.71% to transcription (K) (Table 1).

Fig. 2
figure 2

The graphical circular genomic map of LJBV19 using the CGview server. Circles represented, from inner to outer: scale marks; GC skew (green, positive skew; purple, negative skew); GC content; reverse COG annotated coding sequences; protein-coding genes on reverse strand; protein-coding genes on forward strand; forward COG annotated coding sequences

Full size image
Table 1 COG categories of coding proteins in the B. velezensis LJBV19 genome
Full size table

CAZymes analysis

There were 138 putative CAZymes-coding genes in the LJBV19 genome, including 48 glycoside hydrolases (GHs), 40 glycosyl transferases (GTs), 3 polysaccharide lyases (PLs), 26 carbohydrate esterases (CEs), 7 auxiliary activities (AAs), and 14 carbohydrate-binding modules (CBMs) (Fig. 3A). Moreover, 13 (9.42%) CAZymes with amino-terminal signal peptides for guiding through cytoplasmic membrane as secreted enzymes were crucial for LJBV19 biological activity (Chen et al. 2021). The genome of LJBV19 had 5, 4, 3, and 1 secreted CAZymes in the GHs, CEs, CBMs, and PL families, respectively (Fig. 3A). CAZymes degrade plant polysaccharides by enzymatic reaction (Chen et al. 2021). There were genes encoding for possible antifungal CAZymes, including 6-phospho-β-galactosidase (GH1), 6-phospho-glucosidase (GH4), endo-1,4-β-glucanase (GH5), β-glucanase (GH16), and endoglucanase (GH51) for cellulose degradation, chitinase (GH18), and chitosanase (GH46) (Fig. 3B). The functions of annotated genes involved in hydrolases were validated at the metabolic level, showing that LJBV19 could produce cellulose (0.53 ± 0.00 U/mL), chitinase (0.14 ± 0.01 U/mL), and chitosanase (0.11 ± 0.01 U/mL). These CAZymes in the genome LJBV19 can degrade the cell wall components of pathogens, which played an important role in fungal antagonism (Shafi et al. 2017). For example, β-chitinase or glucanase had the ability to inhibit infection by B. cinerea and C. gloeosporioides (Hamaoka et al. 2021).

Fig. 3
figure 3

Distribution of carbohydrate active enzymes (CAZymes) families in the genome of B. velezensis LJBV19. A The classification of CAZymes in the LJBV19 genome. B Functional characterization of glycoside hydrolase family

Full size image

Comparative genomics analysis

Comparative analysis among the genome sequences of three closely related strains B. velezensis FZB42, B. amyloliquefaciens DSM7T, and B. subtilis 168T with the LJBV19 were performed (Table 2). Genome features of the four strains were annotated based on NCBI to ensure the same annotation conditions. Comparative results revealed that the genome size of LJBV19 (3,973,013 bp) was similar to DSM7T (3,980,199 bp) and FZB42 (3,918,596 bp), but smaller than 168T (4,215,606 bp). The GC content of LJBV19 (46.50%) was the same as FZB42 (46.50%), was approximately equal to DSM7T (46.1%), and was higher than 168T (43.5%). In addition, there was no plasmid in the four Bacillus genomes.

Table 2 Genomic features of B. velezensis LJBV19 and comparison with B. velezensis FZB42, B. amyloliquefaciens DSM7T, and B. subtilis 168T
Full size table

To evaluate the evolutionary distance among the four strains, their whole genome sequences were compared by Mauve program with default parameters (Fig. 4A). The alignments revealed no significant insertion of large regions or large local collinear block (LCB) inversion between LJBV19 and FZB42. Compared to DSM7T and 168T, a number of gene insertions or deletions and LCB inversions were present in LJBV19. More LCB inversions were occurred when LJBV19 compared with 168T showing that the LJBV19 genome was more similar to DSM7T than to 168T. The synteny plot of the pairwise alignments from Mauve program was consistent with taxonomic position of LJBV19.

Fig. 4
figure 4

Comparison of B. velezensis LJBV19 genome sequences against B. velezensis FZB42, B. amyloliquefaciens DSM7T, and B. subtilis 168T. A Mauve progressive alignment of the LJBV19, FZB42, DSM7T, and 168T. LJBV19 genome was used as the reference. Boxes with the same color indicated syntenic regions, and colored lines connected homologous regions. Boxes above the center line were forward regions and below the center line were reverse regions. The scale was in nucleotides. B Venn diagram showing the numbers of shared and unique clusters of orthologous genes

Full size image

LJBV19 genome sequences were compared with above three genome sequences in order to identify the specific genes of LJBV19 (Fig. 4B). There were 1199 conserved genes shared with LJBV19, FZB42, DSM7T, and 168T. Comparison of orthologous genes showed that there were 2956 genes in common with average 85.51% identity between LJBV19 and FZB42, 2709 genes in common with average 78.36% identity between LJBV19 and DSM7T, and 1527 genes in common with average 44.17% identity between LJBV19 and 168T. Moreover, a total of 179 unique genes were present in the genome of LJBV19, and the functions of most of these genes need further confirmation. The result showed that the four strains had a conserved genomic structure and genetic homogeneity with some inversion events during evolution.

Comparison of gene clusters related to secondary metabolites

Bacillus species can secrete secondary metabolites with broad biological activities, such as antimicrobial, antiviral, and nematocidal action, protecting the plant against pathogens (Keswani et al. 2020). There were 13 gene clusters involved in the synthesis of secondary metabolites in the LJBV19, covering 18.93% (752.05 kb) of its genome (Table 3; Fig. S4). These gene clusters were consisted of three non-ribosomal peptide synthetase (NRPS) clusters, three trans-acyl transferase polyketide synthetase (transAT-PKS) clusters, two terpene clusters, one other unspecified ribosomally synthesized and post-translationally modified peptide (RiPP-like) cluster, one type 3 polyketide synthetase (T3PKS) cluster, one lanthipeptide-class-ii cluster, and one “other” type of gene cluster. Eight clusters corresponding with the production of identified secondary metabolites, including surfactin, butirosin, macrolactin H, bacillaene, fengycin, difficidin, bacillibactin, and bacilysin, matched to 82, 7, 100, 100, 100, 100, 100, and 100% of the known gene clusters, respectively. Moreover, five gene clusters with a total length of 116.62 kb encoding potential novel secondary metabolite-related proteins without previously known description.

Table 3 Comparative analysis of secondary metabolite clusters of B. velezensis LJBV19 with reference genomes
Full size table

The locations and products of secondary metabolite gene clusters in LJBV19, FZB42, DSM7T, and 168T were compared (Fig. 5). Interestingly, the core biosynthetic genes in the four strains were similar and the products of core genes exhibited very high homologues at the amino acid level (Fig. 6). The results showed that eight (clusters 1, 4, 6, 7, 8, 9, 11, and 13) involved in the biosynthesis of secondary metabolites in LJBV19 also existed in FZB42, DSM7T, and 168T strains. Among the eight clusters, five gene clusters were identified and specifically involved in the synthesis of surfactin (srfAA, srfAB, srfAC), bacillaene (baeCDEGJLMNR), fengycin (fen, myc, and yng), bacillibactin (dhbF), and bacilysin (bacD). However, B. amyloliquefaciens DSM7T lacked the genes fenA, fenB, and fenC for fengycin biosynthesis, whereas B. subtilis 168T lacked fenF, mycA, mycB, and mycC compared to B. velezensis LJBV19, suggesting that LJBV19 with the ability of fengycin synthesis may have stronger antimicrobial activity than DSM7T and 168T. Surfactin, bacillaene, fengycin, bacillibactin, and bacilysin were also observed in fungal antagonistic Bacillus spp. and proved to be antimicrobial compounds (Ravi et al. 2021). Surfactin and fengycin as non-ribosomal synthesis of cyclic lipopeptides (cLPs) have been proved to enhance plant defense response to pathogens. For example, the supernatant with surfactin and fengycin produced by B. subtilis GLB191 had direct antifungal effect and induced plant defense response, protecting grape against Plasmopara viticola (Li et al. 2019). Surfactin and fengycin exhibited antimicrobial and antiviral activities by altering membrane integrity of pathogens (Chen et al. 2015; Olishevska et al. 2019). Moreover, surfactin was related to quorum-sensing, biofilm formation, and root colonization (Anckaert et al. 2021). Bacillaene, a linear molecule with two amide bonds, was synthesized by PKS and selectively inhibited protein biosynthesis of prokaryotes (Moldenhauer et al. 2010). Siderophore bacillibactin with higher affinity for ferric ion possessed antimicrobial properties through depriving essential iron to alter the fitness and aggressiveness of pathogens (Khan et al. 2018). For example, B. velezensis FZB42 with the ability of bacillibactin production inhibited the growth of phytopathogens (Rabbee et al. 2019). Bacilysin, a dipeptide influencing biosynthesis of microbial cell wall by inhibiting the glucosamine-6-phosphate synthase, showed a broad range of antagonistic activity against phytopathogens, such as F. oxysporum, Erwinia amylovora, and Microcystis aeruginosa (Rabbee et al. 2019; Nannan et al. 2021). The other three (clusters 4, 8, and 9) were involved in unknown secondary metabolite-related proteins encoding terpene, terpene, and T3PKS, respectively.

Fig. 5
figure 5

Comparation of the location and products of secondary metabolite gene clusters in B. velezensis LJBV19, B. velezensis FZB42, B. amyloliquefaciens DSM7T and B. subtilis 168T

Full size image
Fig. 6
figure 6

Comparisons of secondary metabolite clusters in LJBV19, FZB42, DSM7T, and 168T. Red, pink, blue, green, and gray indicated core biosynthetic genes, additional biosynthetic genes, transport-related genes, regulatory genes, and other genes, respectively. The core biosynthetic genes (red) were marked

Full size image

Cluster 3 present in three strains excluding 168T encoding PKS-like was involved in the synthesis of butirosin. Butirosin was as a 2-deoxystreptamine (DOS)-containing aminoglycoside antibiotic, and the key part of biosynthetic gene clusters (ydhFR and pksF) was identified in B. circulans SANK 72073 (Kudo et al. 2005). Importantly, cluster 3 had only 7% similarity with known gene cluster of butirosin indicating that B. velezensis LJBV19 may be able to produce a structurally novel antibiotic compound. Clusters 5 and 10 were shared between LJBV19 and FZB42 encoding transAT-PKS. Cluster 5 had the core genes pks2A, pks2B, pks2C, pks2D, pks2E, pks2F, and pks2G related to the biosynthesis of macrolactin H (Fig. 6). Macrolactin belonging to nonribosomal synthesis of polyketides exhibited antimicrobial activity against gram-negative (i.e., Escherichia coli) and gram-positive (i.e., Staphylococcus aureus) bacteria and fungi (i.e., C. gloeosporioides, B. cinerea, and R. solani) (Ton That Huu et al. 2021). The genes difAFJHIJKL involved in the synthesis of difficidin were found in cluster 10 (Fig. 6). Difficidin, an antibaterial polyketides synthesized by NRPS, showed a broad spectrum of activity against aerobic (i.e., Pseudomonas aeruginosa and Salmonella typhimurium) and anaerobic (i.e., Clostridium difficile) bacteria (Rabbee et al. 2019). Two clusters (clusters 2 and 12) were only present in strain LJBV19: clusters 2 and 12 were shown to direct unknown compounds, encoding RiPP-like and class II lantipeptide, respectively. RiPP with a wide variety of structural features had antifungal, antibacterial, and antiviral activities (Ortega and van der Donk 2016). For example, thioamitides belonging to RiPP caused mitochondrial dysfunction and triggered apoptosis by inhibiting mitochodrial ATP synthase of pathogenic microorganisms (Eyles et al. 2021). Lantipeptide as antibacterial peptides synthesized by the ribosome destroyed the cell walls or membranes of pathogens, resulting in the outflow of small molecules and the dissipation of membrane potential (Dufour et al. 2007). For example, Streptomyces griseus S4-7 producing a class II lantipeptide called grisin suppressed wilt of strawberry caused by F. oxysporum (Kim et al. 2019). These results revealed that the secondary metabolite clusters in the LJBV19 genome were highly similar to gene clusters in fungal antagonistic strain B. velezensis FZB42, which was one of most important commercially available agents, such as RhizoVital®, RhizoPlus®, and Taegro® (Rabbee et al. 2019). Moreover, two unique clusters in LJBV19 genome would encode potential novel metabolites with unknown description. All results indicated that LJBV19 was expected to become a natural antagonist of plant pathogens.

Conclusions

Bacillus velezensis LJBV19, isolated for grapevine rhizosphere soil, showed a broad-spectrum antimicrobial activity against 12 plant pathogens. Whole genome sequencing, annotation, and genomic analysis revealed the structure and function of LJBV19 genome. Among 3993 ORFs in LJBV19 genome, 3308 ORFs were classified into 19 COG categories, such as secondary metabolites biosynthesis, transport and catabolism (Q), and carbohydrate transport and metabolism (G). Hydrolases were predicted by CAZy database and validated at the metabolic level, including cellulose (0.53 ± 0.00 U/mL), chitinase (0.14 ± 0.01 U/mL), and chitosanase (0.11 ± 0.01 U/mL). There were 13 gene clusters related to the biosynthesis of secondary metabolites in LJBV19 genome. Comparative genomic analysis confirmed the taxonomic position of LJBV19 and two unique clusters (clusters 2 and 12) in strain LJBV19. Taken together, these findings showed that LJBV19 possessed the necessary genetic machinery as fungal antagonistic agent and promoted its application.