Introduction

Microbial diseases are one of the main causes of reduced productivity of many agricultural crops (Vieira et al. 2006) and the control of these diseases is usually carried out by combining crop management and use of pesticides and resistant cultivars (Jayaswal et al. 1990; Vieira et al. 2006). The application of pesticides is the primary control method, but their use is controversial because of harmful effects on the environment and the selection of resistant pathogens (Kim et al. 1999; Yu et al. 2014). There are also microbial diseases for which there are still no products with proven effectiveness (Vieira et al. 2006).

The use of microorganisms that are able to protect plants against phytopathogens has emerged as a more rational and safe alternative to traditional methods (Lucy et al. 2004; Hong et al. 2016). Endophytic bacteria colonize the interior of the host plant without causing apparent damage (Hallmann et al. 1997; Lopes et al. 2017). These bacteria can provide many benefits to plants, including growth promotion (Barka et al. 2002; Tanuja et al. 2013), induction of systemic resistance (Bakker et al. 2007), nitrogen supply by biological fixation (Jha and Kumar 2007), increased tolerance to environmental stresses (Andreolli et al. 2013), and protection against phytopathogenic microorganisms (Coombs et al. 2004; Hong et al. 2015), insects, nematodes (D’Alessandro et al. 2014) and herbivory (Sullivan et al. 2007). In this context, endophytic bacteria are ideal for the biological control of plant diseases because they often occupy the same habitat that phytopathogenic microorganisms and inhibits them without causing damage to the host plant and the environment (Melnick et al. 2008).

Bacillus species has been shown to be a potential source of several antimicrobials, such as ribosomally synthesized antimicrobial peptides (bacteriocins) (Abriouel et al. 2011), non-ribosomally synthesized peptides (NRPs), polyketides (PKs), as well as uncommon antimicrobials and other secondary metabolites (Fickers 2012; Lopes 2017). Bacillus subtilis, Bacillus amyloliquefaciens, Bacillus licheniformis, Bacillus cereus and Bacillus thuringiensis are the best studied species for antimicrobials production within the genus (Mondol et al. 2013). Although studies have shown that endophytic B. thuringiensis and B. amyloliquefaciens can increase protection against phytopathogens and promote plant growth (Chen et al. 2007; Jeong et al. 2016), and that B. thuringiensis can still increase seed germination (Tanuja et al. 2013) and root nodulation (Mishra et al. 2009; Tanuja et al. 2013), the association between Bacillus toyonensis and plants was not investigated yet.

Another way to control microbial diseases in plants is by inhibiting the quorum sensing of phytopathogenic bacteria. Quorum sensing is population density-dependent mechanism for bacterial communication and gene regulation (Fuqua et al. 2001), by which pathogenic bacteria, especially Gram-negative, can modulate the production of virulence factors (Von Bodman et al. 2003). Bacillus has demonstrated the ability to produce quorum quenching enzymes which hydrolyze, such as N-acyl homoserine lactonase (AHLase), or modify the chemical structure of quorum sensing signaling molecules in Gram-negative bacteria and have variable substrate spectra (Chen et al. 2013). Studies indicate that the use of quorum sensing inhibitors is promising, thus expanding the current methods to control bacterial diseases in plants.

In this study, the whole-genome sequenced endophytic B. toyonensis BAC3151, which has demonstrated antimicrobial activity and quorum sensing inhibition of phytopathogenic bacteria (Lopes et al. 2015), was analyzed in silico to determine the genes involved in the synthesis of potential antimicrobial compounds. To our knowledge, this is the first comprehensive analysis focusing on the potential of B. toyonensis for the production of antimicrobials and provides valuable insights into how the strain BAC3151 can be used for the development of new strategies to control microbial diseases in plants.

Materials and methods

Bacterial strain and growth conditions

A strain identified initially as Bacillus thuringiensis BAC3151 by 16S rRNA gene sequencing was previously isolated from the leaves of the common bean (Phaseolus vulgaris) (Costa et al. 2012) and reclassified as Bacillus toyonensis BAC3151 by means of pairwise genome calculations of the average nucleotide identity (ANI) (Federhen et al. 2016). This strain is part of the bacterial strain collection of the Laboratory of Microorganism Molecular Genetics/BIOAGRO of the Federal University of Viçosa (Universidade Federal de Viçosa—UFV). B. toyonensis BAC3151 was cultured in 10% tryptone soy agar (TSA) (1.5 g/l tryptone, 0.5 g/l soy peptone, 1.5 g/l NaCl, pH 7.3) at 28 °C for the DNA isolation.

DNA isolation and genome sequencing

For genome sequencing, total DNA was isolated from culture grown for 16 h using the Wizard® genomic DNA purification kit (Promega) according to the manufacturer’s instructions. The total DNA was sequenced by Macrogen Inc. (Seoul, South Korea) using an Illumina HiSeq 2000 platform (100 bp paired-end). A total of 21,512,378 high-quality reads and 2,172,750,178 high-quality bases were generated, producing ~355× coverage of the genome.

Data processing and assembly

The raw data obtained from sequencing were processed using the Phred-Phrap-Consed package (Rasko et al. 2008). The quality value used for the base-calling program was Q ≥ 20. In the next step, 112 contigs were initially assembled using SOAPdenovo version 2.4 (Luo et al. 2012) and those < 500 bp were removed, resulting in a total of 78 contigs. The chromosome and plasmids contigs were then separated by Blastn. The chromosome contigs were assembled in the correct order and orientation using the B. toyonensis BCT-7112 genome as reference on the program Mauve version 2.3.1 (Darling et al. 2010) with default parameters.

Genome annotation

The annotation process involved the use of various algorithms in a multistep procedure. The structural annotation was performed using the following software: FgenesB: gene predictor (Solovyev and Salamov 2011), RNAmmer: rRNA predictor (Lagesen et al. 2007), and tRNAscan-SE: tRNA predictor (Lowe and Eddy 1997). The functional annotation was performed by blast analyses (Altschul et al. 1990) with the Non-Redundant (NR) database, Conserved Domains Database (CDD), Gene Ontology (GO), Swiss-Prot and Clusters of Orthologous Groups (COG), and InterProScan analysis (Jones et al. 2014) (e-value < 10−5). Annotation was also performed using NCBI Prokaryotic Genomes Annotation Pipeline (PGAP) (Angiuoli et al. 2008).

Phylogenomic analysis

The phylogenomic analysis of B. toyonensis BAC3151 was performed using 699 core genes of 54 Bacillus strains deposited at NCBI (http://www.ncbi.nlm.nih.gov/genome/browse/). In the process, the core genes were aligned using Muscle version 3.8.31 (Edgar 2004), and non-matching regions in the alignment were masked and then removed using Gblocks (Castresana 2000). The remaining regions from all of the alignments were compiled to form an alignment of large sequences, which was then used as input for the program Phylip v. 3.2 (Felsenstein 1989). A tree was constructed using the neighbor-joining method with 1000 replicates.

Comparative analysis among Bacillus strains

Efficient Database framework for comparative Genome Analyses using BLAST score Ratios (EDGAR) (Blom et al. 2009) was used to evaluate the gene sharing among genomes of Bacillus strains available at NCBI (http://www.ncbi.nlm.nih.gov/genome/browse/). To analyze putative secondary metabolite gene clusters and bacteriocins, were used antiSMASH (Blin et al. 2017) and BAGEL3 (van Heel et al. 2013) software, respectively, with their default parameters. Compounds with antimicrobial activity were also mined from the published literature, their sequences were obtained from GenBank (https://www.ncbi.nlm.nih.gov/protein/) and homologous sequences were extracted from the analyzed genomes by similarity using Blastp (e-value < 10−5; identity ≥ 85%; length ≥ 90%). The ClustalW algorithm (Larkin et al. 2007) was used to align the sequences.

Nucleotide sequence accession numbers

This Whole Genome Shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession LDKD00000000. The version described in this paper is version LDKD02000000.

Results

General genomic features of Bacillus toyonensis BAC3151

The strain BAC3151 was first classified into B. thuringiensis based on a 16S rRNA gene sequence analysis (Costa et al. 2012). However, the whole-genome sequenced strain was 99.493% identical by ANI to the type genome of B. toyonensis (strain BCT-7112), and thus it was reclassified as B. toyonensis. The draft genome sequence of B. toyonensis BAC3151 has a total length of 5,740,808 bp (1 chromosome scaffold and 37 plasmid contigs), an average G + C content of 34.9%, and an N50 length of 299,415. A total of 5,706 protein-coding DNA sequences (CDSs) was predicted, comprising 4,685,804 bp, which resulted in a protein coding percentage of approximately 81.6%. The G + C content of the CDSs was slightly higher than the total G + C content, at 35.8%. The gene density was 0.993 CDS/kb, with a mean size of 821 bp per CDS. The draft genome also had 46 RNAs genes (40 tRNAs, 1 rRNA and 5 non-coding RNAs) and 303 pseudogenes. General features of the B. toyonensis genomes are summarized in Table 1. The strains were found to have similar mean G + C content, coding percentage, length and density of CDSs. However, there were significant differences in gene number and strain-specific genes.

Table 1 Genomic features of Bacillus toyonensis strains
Full size table

Phylogenomic analysis

The phylogenomic tree inferred from 699 core gene of 54 Bacillus strains (including 18 B. thuringiensis strains, 12 B. cereus strains, 8 Bacillus anthracis strains, 4 B. toyonensis strains, 4 B. amyloliquefaciens strains, 4 B. subtilis strains, 1 Bacillus velezensis strain, 1 B. licheniformis strain, 1 Bacillus clausii strain, and 1 Bacillus sp. strain) is presented in Fig. 1. The cluster B. cereus sensu lato containing the B. toyonensis, B. thuringiensis, B. anthracis and B. cereus strains apart from other Bacillus members under study can be recognized. The endophytic B. toyonensis BAC3151 had a closer relationship with the insect strain B. toyonensis VU-DES13. B. thuringiensis KB1 grouped together with B. cereus strains, while B. velezensis LS69 and Bacillus sp. Pc3 were closely related with B. amyloliquefaciens strains.

Fig. 1
figure 1

Phylogenomic tree of 54 Bacillus strains based on 699 core genes. The tree was constructed with Phylip (Felsenstein 1989) using the neighbor-joining method with 1000 replicates. Listeria monocytogenes EGD-e was used as outgroup. The bootstrap values are shown at the nodes. The scale bar shows five nucleotide substitutions per 1000 nucleotides. Green boxes highlight strains with reported antimicrobial activity; black boxes indicate strains with high insecticidal or nematicidal activity. Green squares show strains isolated from plants; black, insect; brown, soil; yellow, food; red, human or animal; gray, other sources or sources not specified. The accession numbers of the sequences are given in parentheses

Full size image

Gene sharing among Bacillus strains

B. toyonensis shared more orthologous genes with B. thuringiensis than with other Bacillus species (Fig. 2a). Considering the four genomes of B. toyonensis strains (BacAer, BAC3151, VU-DES13 and BCT-71-12) available until August 2017 at NCBI, were identified 3811 CDSs shared between them (Fig. 2b), which correspond to 65.5 to 93.0% of the CDSs of the strains. BAC3151 had more CDSs in common with VU-DES13 (4973 CDSs), and fewer with BacAer (4243 CDSs). CDSs involved in the general physiological and survival functions of the species were well conserved in all four strains.

Fig. 2
figure 2

Venn diagrams showing the sharing CDSs between Bacillus species (a) and B. toyonensis strains (b). Overlapping regions represent CDSs that are common between the genomes. Values outside of the overlapping regions indicate the number of CDSs in each genome without orthologs in the other genomes

Full size image

The total number of CDSs of all B. toyonensis strains was 6,980. Furthermore, 245 strain-specific CDSs in BCT-7112, 250 in VU-DES13, 416 in BAC3151 and 704 in BacAer were found. Most of the strain-specific CDSs were associated with hypothetical proteins, while several others were involved in response to environmental stresses, antimicrobials biosynthesis, processes of pathogenicity and virulence for animal (in VU-DES13 and, mainly, BacAer), and drug resistance. Among B. toyonensis strains, BAC3151 had the highest number of genes involved in microbial antagonism and was the only presenting, e.g., gene clusters of thuricin-like, uncharacterized linear azole-containing peptide (LAP), and lactococcin 972 family bacteriocin in the same strain (Fig. 3).

Fig. 3
figure 3

Sets of antagonism genes present in Bacillus strains that show antimicrobial activity and/or strains of agricultural importance

Full size image

Putative genes of microbial antagonism found in B. toyonensis BAC3151

Considering the four available genomes of B. toyonensis strains, the species had a higher frequency of bacteriocin gene clusters than the most reported Bacillus species for antimicrobials production (Table 2). In addition, putative biosynthetic genes of several other antimicrobial compounds, including NRPs, AHLase, chitinases and novel bacteriocins, were identified in B. toyonensis BAC3151. The strain also showed a gene repertoire involved in antagonist greater than that of other Bacillus strains with efficient antimicrobial activity, such as B. thuringiensis KB1 and B. subtilis fmb60 (Jeong et al. 2016; Yang et al. 2016) (Fig. 3, Table 3).

Table 2 Number of putative bacteriocin gene clusters identified in Bacillus genomes
Full size table
Table 3 Biosynthetic gene clusters and individual genes involved in antagonist found in Bacillus species with antimicrobial activity
Full size table

Bacteriocin gene clusters found in BAC3151 included class I and class II bacteriocins. Class I bacteriocins are small peptides (< 10 kDa) ribosomally synthesized, undergoing posttranslational modifications. According to the modification differences, this class can be subdivided into different subclasses (Arnison et al. 2013). Three class I bacteriocins gene clusters were identified in BAC3151, encoding: one lanthipeptide similar to thuricin, one paeninodin-like (lasso peptide) and one LAP. Other gene clusters encoded two class II bacteriocins (unmodified bacteriocins): BhlA-like and lactococcin 972 family bacteriocin. The gene clusters of lactococcin 972 family bacteriocin and thuricin-like (Fig. 4a, b) are in plasmids, while all other bacteriocin gene clusters are on the chromosome (Table 3). Class III bactericin gene clusters (encoding large bacteriocins, > 10 kDa) (Zhao and Kuipers 2016; Turano et al. 2017) were not identified in BAC3151.

Fig. 4
figure 4

Biosynthetic gene clusters of putative antimicrobials of B. toyonensis. a Schematic representation of the previously characterized thuricin gene cluster of B. thuringiensis As 1.1013 (GenBank accession: KJ504104.1) and thuricin-like lanthipeptide gene cluster of B. toyonensis BAC3151. b Lactococcin 972 gene cluster of Lactococcus lactis IPLA 972 (GenBank accession: AJ002203.2) and lactococcin 972 family bacteriocin gene cluster of B. toyonensis BAC3151. Gene clusters of petrobactin (c) and bacillibactin (d) found in Bacillus species. Red, precursor peptide genes; blue, genes involved in biosynthesis/modification; yellow, transport/immunity; gray, other genes

Full size image

The biosynthetic genes of NRPs (petrobacin and bacillibactin siderophores, Fig. 4c, d, respectively), chitinases and AHLase in BAC3151 can be important for the inhibition of phytopathogens by competition for iron, degradation of fungal chitin and inhibition of bacterial quorum sensing, respectively. For AHLase, it was also observed that the motif 106HXDH-H169-D191-Y194 necessary for the enzyme activity (Dong et al. 2000, 2002; Lu et al. 2006) was conserved in BAC3151 and other Bacillus species (Fig. 5).

Fig. 5
figure 5

Analysis of AHLases from Bacillus sp. a Phylogenetic tree of AHLase genes generated on the basis of the neighbor-joining method with 1000 repetitions using MEGA 7.0. Bootstrap values > 50 are shown. The GenBank accession numbers of the sequences are indicated in parentheses. b Multiple alignment of predicted AHLases of various Bacillus species. Gray shading indicates sites that were not completely conserved. The amino acid residues necessary for AHL-lactonase activity are boxed. Bt, Bacillus toyonensis; Bw, Bacillus weihenstephanensis; Bth, Bacillus thuringiensis; Bc, Bacillus cereus; Bwi, Bacillus wiedmannii; Bs, Bacillus subtilis

Full size image

Discussion

The Bacillus cereus group (or Bacillus cereus sensu lato) comprises eight closely relatated species, including Bacillus cereus, Bacillus thuringiensis, Bacillus anthracis, Bacillus toyonensis, Bacillus mycoides, Bacillus pseudomycoides, Bacillus weihenstephanensis and Bacillus cytotoxicus (Helgason et al. 2000; Zwick et al. 2012; Guinebretière et al. 2013; Jiménez et al. 2013). Traditionally, the extensive similarity between them has been revealed by different techniques, such as 16S rRNA gene sequencing (Bavykin et al. 2004), multilocus sequence typing (MLST) (Priest et al. 2004), genomic mapping (Carlson et al. 1996), pulsed-field gel electrophoresis (PFGE) (Carlson et al. 1994), multilocus enzyme electrophoresis (MEE) (Helgason et al. 2000), variable number tandem repeat (VNTR) mapping, BOX-PCR fingerprinting (Kim et al. 2002), amplified fragment length polymorphisms (AFLP) (Ticknor et al. 2001), and, currently, genome sequencing (Zwick et al. 2012; Böhm et al. 2015).

B. toyonensis BCT-7112, which has been used as active ingredient of the feed additive preparation Toyocerin, was primarily identified as Bacillus cereus. However, the strain showed significant genomic differences from the type strains of the B. cereus group (ANI < 92%) and it was reclassified as a novel species (Jiménez et al. 2013). B. toyonensis, B. pseudomycoides and B. cytotoxicus have been distinguishable at species level (ANI > 96%), while other species within the B. cereus group, as observed in the phylogenetic analysis of this study (Fig. 1), do not match individual phylogenic groups (Böhm et al. 2015). For the indistinguishable species by ANI, differences in phenotype and pathological effects, are mainly resulted from plasmid-mediated characteristics rather than in chromosome in many cases (Helgason et al. 2000; Böhm et al. 2015).

Differences in the number of genes and strain-specific genes, as we observed for B. toyonensis strains (Table 1; Fig. 2b), can be primarily explained by mobile genetic elements, such as plasmids, transposons, phages, integrons and genomic islands (Frost et al. 2005; Fang et al. 2011; Böhm et al. 2015). Moreover, the large number of hypothetical proteins identified in B. toyonensis strains may also be the result of horizontal gene transfer, including genes from phylogenetically distant organisms for which there is currently no molecular characterization. Indeed, Bacillus can easily acquire genetic material from other bacteria (Gonzalez et al. 1982; Haack et al. 1996; Vilas-Bôas et al. 1998; Böhm et al. 2015) and the acquisition of foreign genes can contribute to improved adaptation of populations to specific niches.

In the environment, microorganisms compete for limited resources within a community and competition mechanisms are characterized by antagonistic effects. Direct microbial antagonism occurs by the production of compounds that kill or inhibit the growth of other microorganisms, while indirect antagonism occurs by competition for nutrients and space. Several Bacillus species have shown activity against other microorganisms (Jeong et al. 2016; Yang et al. 2016; Liu et al. 2017). We previously reported that B. toyonensis BAC3151 had antimicrobial activity against Gram-negative and Gram-positive bacteria, including important phytopathogenic bacteria (Lopes et al. 2015). In this study, the genome mining revealed that the strain has potential to produce a variety of antimicrobials, mainly bacteriocins.

Bacteriocins found in Bacillus are becoming increasingly important due to their spectrum of inhibition, which is sometimes broader than many lactic acid bacteria bacteriocins and may include Gram-negative bacteria, yeast and filamentous fungi, in addition to Gram-positive bacteria (Abriouel et al. 2011). To date, bacteriocins of the B. cereus group have been identified and characterized mainly in B. cereus and B. thuringiensis strains.

Class I bacteriocins belonging to lantipeptide subclass are peptides containing unusual amino acids, such as dehydroalanine/dehydrobutyrine and lanthionine/methyl-lanthionine residues, introduced by posttranslational modifications (Arnison et al. 2013). While class I lanthipeptides are modified by two distinct enzymes (dehydratase LanB and cyclase LanC), class II lanthipeptides are modified by a bifunctional lanthionine-introducing enzyme (LanM) (Knerr ànd van der Donk 2012). Lanthipeptides with antimicrobial activity are named lantibiotics (McAuliffe et at. 2001). Many lantibiotics exhibit broad-spectrum antimicrobial activity against clinical Gram-positive pathogens, including multidrug-resistant strains (Dischinger et al. 2014; Sandiford 2014). So far, reports about various bacteriocins in the B. cereus group have been published, but few were about lantibiotics. Some studies suggested that the B. cereus group may contain unidentified lanthipeptide gene clusters (Wang et al. 2014; Xin et al. 2015). In BAC3151, one thuricin-like lanthipeptide gene cluster was found in plasmid. Thuricins are typically produced by B. thuringiensis and may be encoded on chromosome or large plasmids (Favret and Yousten 1989; Murphy et al. 2011), but thuricin-like clusters have also been described in other species (Ugras et al. 2013). These putative clusters have not yet been experimentally verified and may be a source of novel lantibiotics.

Class I bacteriocins of the lassopeptide subclass have an N-terminal macrolactam ring that is threaded by the C-terminal tail resulting in a unique lasso structure isolated from proteo- and actinobacterial sources (Hegemann et al. 2015). In general, lasso peptide synthesis requires at least three genes, encoding a precursor peptide A, a cysteine protease B, and an ATP-dependent lactam synthetase C (Maksimov et al. 2012). Known lasso peptides display antimicrobial activity by enzyme inhibition, and microcin J25 produced by Escherichia coli AY25 has served as a model for studies (Delgado et al. 2001; Mukhopadhyay et al. 2004). There is one paeninodin-like lasso peptide gene cluster in BAC3151. The paeninodin is a novel lasso peptide recently described from Paenibacillus dendritiformis and has potential antimicrobial activity, although this activity has not yet been confirmed (Zhu et al. 2016).

Class I bacteriocins of the LAP subclass have a distinguishing heterocyclic ring of oxazoles and thiazoles (Melby et al. 2011; Banala et al. 2013). Microcin B17 produced by E. coli and streptolysin S produced by lactic acid bacteria are model of representive LAPs (Nizet et al. 2000; Heddle et al. 2001; Cox et al. 2015). LAP subclass has been extended with plantazolicin A and B produced by B. amyloliquefaciens FZB42, and they have a unknown mechanism of action against other Bacillus, in addition to being associated with nematicidal activity (Banala et al. 2013; Liu et al. 2013). Only these two LAPs have been characterized in Bacillus before. However, a putative gene cluster of LAP was identified in BAC3151, meaning that novel LAPs can be found in B. toyonensis.

Class II bacteriocins present in BAC3151 were lactococcin 972 family bacteriocin and BhlA-like. Lactococcin 972 is plasmid-encoded and typically produced by Lactococcus lactis (Martı́nez et al. 1999). The lactococcin 972 gene cluster comprises structural and immunity/transport genes (Martı́nez et al. 1999; Sánchez et al. 2000) (Fig. 4b). This bacteriocin has a potent antimicrobial activity against lactococci and Lactobacillus (Martı́nez et al. 1995; Martı́nez et al. 1996). Unlike others class II bacteriocins, its primary target is not the cytoplasmic membrane. Instead, lactococcin 972 binds to the cell wall precursor lipid II and inhibits septum biosynthesis (Martı́nez et al. 2008). To our knowledge, similar bacteriocin has not yet been reported in Bacillus species. On the other hand, genes coding for BhlA have been described in Bacillus pumilus and B. licheniformis, and structural analysis of their sequence revealed features similar to holins (Kyogoku and Sekiguchi 1996; Aunpad and Panbangred 2012), which are phage-encoded proteins involved in the disruption of cytoplasmic membrane (Ziedaite et al. 2005). The peptide BhlA from Bacillus licheniformis showed antibacterial activity against Gram-positive bacteria, including methicillin-resistant Staphylococcus aureus (MRSA) and Micrococcus luteus, by destroying cell membrane (Anthony et al. 2010).

We also found gene clusters putatively encoding NRPs petrobacin and bacillibactin in BAC3151. Petrobactin and bacillibactin are catecholate siderophores, which are molecules of iron acquisition, encoded by the asb and dhb operons, respectively (Wilson et al. 2006; Chen et al. 2007). Although siderophores are produced by various strains of the B. cereus group, siderophores originating from endophytic bacteria may be especially important for biocontrol in agriculture, given that these bacteria are naturally associated with plants. In this way, iron competition, an essential growth element, could indirectly suppress the growth of phytopathogens, such as fungi, which produce lower affinity siderophores (Kloepper et al. 1980; Compant et al. 2005). In addition, plant-associated bacteria can increase the supply of iron for the plant (Raddadi et al. 2007; Jin et al. 2014).

Another key gene identified was aiiA, which encodes the enzyme AHLase. This enzyme is presumably responsible for the ability of BAC3151 to inhibit the quorum sensing in Gram-negative bacteria, as previously reported by us (Lopes et al. 2015). The observed conservation of the motif 106HXDH-H169-D191-Y194 (Fig. 5) reinforces evidence of AHLase activity from BAC3151. Gram-negative phytopathogenic bacteria use the quorum sensing as a regulatory mechanism for many biological activities, including the production of virulence factors (Von Bodman et al. 2003). The quorum sensing depends on the production, diffusion, and recognition of small signal molecules, usually N-acyl homoserine lactones (AHLs) in Gram-negative bacteria (Fuqua et al. 2001). AHLs share identical homoserine lactone rings, but vary in length and the substitution of the acyl side chain (Whitehead et al. 2001; Zhang 2003). The enzymes AHLases hydrolyse the lactone rings producing molecules that can no longer be used for signalling (Dong et al. 2000; Chen et al. 2013). Thus, the inhibition of the quorum sensing can be an effective strategy to control plant diseases. In addition to using the AHLase producer to suppress the quorum sensing-dependent virulence of phytopathogenic bacteria (Dong et al. 2004), genetically modified plants expressing aiiA can also efficiently quench bacterial quorum sensing and attenuate the virulence (Dong et al. 2001; Quiñones et al. 2005; Ouyang and Li 2016). These findings illustrate the promising potential to explore quorum sensing inhibitors for the control and prevention of infectious plant diseases, especially against bacteria that are not susceptible to treatment with available chemical agents.

With regard to chitinases, they may also be valuable for the control of plant diseases. Chitinases hydrolyse the glycosidic bonds of chitin, an important component of fungal cell wall, leading to cell lysis, and have been demonstrated to successfully inhibit the growth of phytopathogenic fungi. Chitinase-producing bacteria have been reported as biocontrol agents for different kinds of fungal diseases of plants (Kobayashi et al. 2002; Tang et al. 2012, 2017). In the B. cereus group, some strains have shown promising for this purpose. The chitinase-producing endophytic B. cereus strain 65, e.g., was found effective against Rhizoctonia solani in cotton (Pleban et al. 1997). Similarly, B. thuringiensis strains inhibited several fungi, such as Sclerotium rolfsii, Pyricularia grisea, Physalospora piricola, Fusarium gramineum and Fusarium oxysporum (Reyes-Ramirez et al. 2004; Tang et al. 2012, 2017). In addition, chitinases have been used to increase the insecticidal activity of B. thuringiensis, since chitin is a major component of the exoskeleton of insects (Liu et al. 2002; Juárez-Hernández et al. 2015; Ni et al. 2015). Chitinases from B. toyonensis have not yet been characterized and future experimental investigations will be able to analyze their effectiveness for the control of phytopathogenic fungi as well as insect pests.

Conclusions

This study showed that B. toyonensis has a potential not previously reported for the species to produce antimicrobial compounds. Multiple putative genes involved in different types of antagonism are present in the strain BAC3151, including known clusters previously characterized in other bacteria and novel gene clusters. Future studies focusing on the biochemical and structural properties of these antimicrobials may reveal their spectrum of activity and mode of action and to develop new strategies to control pathogenic microorganisms in agriculture or even in other areas.