Kill and cure: genomic phylogeny and bioactivity of Burkholderia gladioli bacteria capable of pathogenic and beneficial lifestyles

Abstract

Burkholderia gladioli is a bacterium with a broad ecology spanning disease in humans, animals and plants, but also encompassing multiple beneficial interactions. It is a plant pathogen, a toxin-producing food-poisoning agent, and causes lung infections in people with cystic fibrosis (CF). Contrasting beneficial traits include antifungal production exploited by insects to protect their eggs, plant protective abilities and antibiotic biosynthesis. We explored the genomic diversity and specialized metabolic potential of 206 B. gladioli strains, phylogenomically defining 5 clades. Historical disease pathovars (pv.) B. gladioli pv. allicola and B. gladioli pv. cocovenenans were distinct, while B. gladioli pv. gladioli and B. gladioli pv. agaricicola were indistinguishable; soft-rot disease and CF infection were conserved across all pathovars. Biosynthetic gene clusters (BGCs) for toxoflavin, caryoynencin and enacyloxin were dispersed across B. gladioli , but bongkrekic acid and gladiolin production were clade-specific. Strikingly, 13 % of CF infection strains characterized were bongkrekic acid-positive, uniquely linking this food-poisoning toxin to this aspect of B. gladioli disease. Mapping the population biology and metabolite production of B. gladioli has shed light on its diverse ecology, and by demonstrating that the antibiotic trimethoprim suppresses bongkrekic acid production, a potential therapeutic strategy to minimize poisoning risk in CF has been identified.

Pan-Genome Analysis Reveals Host-Specific Functional Divergences in Burkholderia gladioli

“…The ANI values were visualized by heatmap.2 function in the gplots package in the R environment ( , accessed on 10 February 2021). We also confirmed the phylogenetic distribution of the 14 B. gladioli in cladistic structure, which were previously well described by Jones et al [ 32 ]. To cover the full diversity of the population, the representative 22 B. gladioli members were selected from different clades of 206 B. gladioli isolates and compared with 14 B. gladioli as in a study of Lin et al [ 39 ].…”

“…The ANI-based phylogenomic relationships also validated in cladistic structure, which were previously well described by Jones et al [ 32 ]. We obtained 4392 concatenated core genes by comparing 14 B. gladioli genomes in this study with the genomes of 22 previously reported ( Figure 2 B).…”

“…The close relationship between human isolates and gladiolus isolates offers a perspective on the origin and acquisition of biological functions that will be necessary for survival or adaptation in a new challenging niche like humans. According to this perspective, studies by Jones et al [ 32 ] and Lipuma et al [ 70 ], suggested that the main source of human-associated Burkholderia is the natural environment. In rice isolates, B. gladioli BSR3 and KACC 18962 were clearly distinguished from strains in gladiolus and human isolates while KACC 18962 shared high ANI scores with FDAARGOS_390 and FDAARGOS_391 of nature isolates (ANI values >99.0%).…”

“…Of secondary metabolites, nonribosomal peptide (NRP), which is synthesized by nonribosomal peptide synthetase (NRPS) working in a ribosome-independent assembly line fashion [ 117 ], has attracted intense interest by its antimicrobial activity [ 118 , 119 ]. In this regard, Burkholderia -produced NRP and NPRS BGCs have been recently explored in some Burkholderia members, including B. gladioli [ 32 , 39 ], B. ambifaria [ 120 ], B. pseudomallei [ 121 ], and B. thailandensis [ 122 ]. Hence, to explore NPRS BGCs of B. gladioli in this study, genome mining was conducted using antiSMASH.…”

“…For comparative analysis, the genomic sequences for B. gladioli strains, including a B. gladioli KACC 11889 whole genome sequence, which was previously developed [ 16 ], were downloaded from the FTP site of Reference Sequence (RefSeq) database at NCBI ( , accessed on 10 February 2021). Subsequently, in selecting genomes for the comparative analyses, we excluded as many as possible bacterial genomes which were isolated from unclear or unshared sites with other strains and were previously studied [ 32 ]. Consequently, fourteen B. gladioli genomes, which were consisted of 4 complete genomes and 10 draft genomes, were carefully included in this study.…”