Putative antibiotic resistance genes present in extant Bacillus licheniformis and Bacillus paralicheniformis strains are probably intrinsic and part of the ancient resistome

Abstract

Whole-genome sequencing and phenotypic testing of 104 strains of Bacillus licheniformis and Bacillus paralicheniformis from a variety of sources and time periods was used to characterize the genetic background and evolution of (putative) antimicrobial resistance mechanisms. Core proteins were identified in draft genomes and a phylogenetic analysis based on single amino acid polymorphisms allowed the species to be separated into two phylogenetically distinct clades with one outlier. Putative antimicrobial resistance genes were identified and mapped. A chromosomal ermD gene was found at the same location in all B. paralichenformis and in 27% of B. licheniformis genomes. Erythromycin resistance correlated very well with the presence of ermD. The putative streptomycin resistance genes, aph and aadK, were found in the chromosome of all strains as adjacent loci. Variations in amino acid sequence did not correlate with streptomycin susceptibility although the species were less susceptible than other Bacillus species. A putative chloramphenicol resistance gene (cat), encoding a novel chloramphenicol acetyltransferase protein was also found in the chromosome of all strains. Strains encoding a truncated CAT protein were sensitive to chloramphenicol. For all four resistance genes, the diversity and genetic context followed the overall phylogenetic relationship. No potentially mobile genetic elements were detected in their vicinity. Moreover, the genes were only distantly related to previously-described cat, aph, aad and erm genes present on mobile genetic elements or in other species. Thus, these genes are suggested to be intrinsic to B. licheniformis and B. paralicheniformis and part of their ancient resistomes. Since there is no evidence supporting horizontal transmission, these genes are not expected to add to the pool of antibiotic resistance elements considered to pose a risk to human or animal health. Whole-genome based phylogenetic and sequence analysis, combined with phenotypic testing, is proposed to be suitable for determining intrinsic resistance and evolutionary relationships.

Fig 1. Whole genome phylogenetic tree.

Number in parentheses; Multi Locus Sequence Type. Clade A corresponds to B. paralicheniformis and clade B to B. licheniformis.

Fig 2. Distribution of CAT protein variants in the phylogenetic tree.

Number in parentheses; Multi Locus Sequence Type. Clade A corresponds to B. paralicheniformis and clade B to B. licheniformis. The colored dots refer to the CAT types shown in Table 1: Variant 1 (blue dots); Variant 2 (pink dots); Variant 3 (amber dots); Variant 4 (red dots); Outliers (green dots).

Fig 3. Distribution of ErmD protein variants in the phylogenetic tree.

Number in parentheses; Multi Locus Sequence Type. Clade A corresponds to B. paralicheniformis and clade B to B. licheniformis. The colored dots refer to the ErmD variants shown in Fig 4: Erm type 1.1 (red dot); Erm type 1.2 (amber dot); Erm type 2 (blue dot); Erm outlier 1 CHCC14814 (green dot); Erm outlier 2 CHCC20375 (purple dot).

Fig 4. Phylogenetic tree of ErmD proteins.

The tree was built using approximately-maximum-likelihood algorithms (Fasttree, see methods) and rooted on the longest branch length. To assess branch length support, Shimodaira-Hasegawa tests were performed in 1,000 resamplings. Each strain code is followed by the gene locus tag number. Three previously published sequences in the ErmD class are included, i.e. L08389 [52], M29832 [53] and M77505 [22]. Sequences of strains to the right of a vertical line are 100% identical. The colored dots designate the ErmD variants: Erm type 1.1 (red dot); Erm type 1.2 (amber dot); Erm type 2 (blue dot); Erm outlier 1 CHCC14814 (green dot); Erm outlier 2 CHCC20375 (purple dot).