Genomic insights into antibiotic resistance and mobilome of lactic acid bacteria and bifidobacteria

Abstract

Lactic acid bacteria (LAB) and Bifidobacterium sp. (bifidobacteria) can carry antimicrobial resistance genes (ARGs), yet data on resistance mechanisms in these bacteria are limited. The aim of our study was to identify the underlying genetic mechanisms of phenotypic resistance in 103 LAB and bifidobacteria using whole-genome sequencing. Sequencing data not only confirmed the presence of 36 acquired ARGs in genomes of 18 strains, but also revealed wide dissemination of intrinsic ARGs. The presence of acquired ARGs on known and novel mobile genetic elements raises the possibility of their horizontal spread. In addition, our data suggest that mutations may be a common mechanism of resistance. Several novel candidate resistance mechanisms were uncovered, providing a basis for further in vitro studies. Overall, 1,314 minimum inhibitory concentrations matched with genotypes in 92.4% of the cases; however, prediction of phenotype based on genotypic data was only partially efficient, especially with respect to aminoglycosides and chloramphenicol. Our study sheds light on resistance mechanisms and their transferability potential in LAB and bifidobacteria, which will be useful for risk assessment analysis.

Figure 1.. Phenotypic resistance profiles of 103 lactic acid bacteria and bifidobacteria.

The minimum inhibitory concentrations (MICs) determined by the microdilution tests and the cut-off MICs that define whether a strain is susceptible or resistant to a particular antibiotic are shown as a heatmap. The names of the strains resistant to five different classes of clinically important antimicrobials are highlighted in red. C, cut-off MICs; FD, feed additive; HM, isolate from human milk or colostrum; M, MICs determined by microdilution tests; NS, isolate of natural microbiota from fermented products (non-starter strain); P, probiotic strain; PC, protective culture; R, resistance; Synercid, quinupristin/dalfopristin; S, starter culture.

Figure 2.. Acquired, intrinsic, and candidate antimicrobial resistance genes (ARGs) found in 103 bacterial strains.

A gene was annotated as an ARG based on the best BLAST hit with a sequence similarity threshold greater than 70%. The intrinsic and acquired nature of ARGs was determined with the aid of mobile genetic element prediction and pan-genome analyses. Candidate (homologous) ARGs were identified based on additional analyses of the hits with lower BLAST similarities (sequence similarity threshold between 40% and 70%).

Figure 3.. Polymorphisms in S12, Lsa(A), and MsrC in 16S and 23S rRNA.

Shown is a section of the sequence alignment in which the mutations presumably associated with resistance are highlighted in red. (A) Substitution of amino acid K43 in S12 was associated with streptomycin resistance. Polymorphisms in 16S rRNA and 23S rRNA in strains of (B) Lactobacillus paragasseri and (C) Lacticaseibacillus rhamnosus confer resistance to different groups of antimicrobials. K-12 and CFT073 represent Escherichia coli strains. (D) Polymorphisms in key motifs of Lsa(A) and homologs were associated with clindamycin resistance (SNPs highlighted in red) and susceptibility (SNPs highlighted in purple). (E) Phylogenetic tree of MsrC protein sequences of E. faecium strains. Shown are the minimum inhibitory concentrations (MICs) of erythromycin. The tree was rooted with an outgroup (E. faecalis IM1312). The exceeded cut-off MICs are shown in bold. MIC, minimum inhibitory concentration; S, susceptible; GEN, gentamicin; KAN, kanamycin; NEO, neomycin; TET, tetracycline; ERY, erythromycin; CHL, chloramphenicol; AMP, ampicillin; QDA, quinupristin/dalfopristin; LIN, linezolid.

Figure 4.. Phenotype–genotype agreement analysis of 103 strains of lactic acid bacteria and bifidobacteria.

In cases where no cut-off minimum inhibitory concentration was defined and in cases where the minimum inhibitory concentration was outside the concentration range of the microdilution test, agreement was not determined (shown in dark grey). ARG, resistance gene; Synercid, quinupristin/dalfopristin.

Figure S1.. Comparison of the genome sequences of different strains showing genomic islands.

B. lactis, Bifidobacterium animalis subsp. lactis.

Figure 5.. Genetic organisation of the detected mobile genetic elements.

(A) Gene tet(M) resides on Tn916. (B) Gene tet(L) is located on an incomplete element that shows sequence similarity to a segment of Tn6079. (C) ANT(6)-Ia, SAT-4, APH(3′)-IIIa, cat, and/or erm(B) are located on elements similar to the enterococcal plasmid pRE25. (D) Gene tet(U) was associated with a putative novel plasmid. Small genomic islands were found in (E) strains of B. animalis subsp. lactis, (F) B. longum IM810, and (G) B. breve IM1386. Probiotic bacterium Limosilactobacillus reuteri IM566 carries plasmids (H) pLR581 and (I) pLR585. (J) Genes tet(S) and ANT(6)-Ia reside on a putative plasmid. (K) Candidate arr-4 is on a putative phage-inducible chromosomal island. Gene function was determined using BLAST and HMMER3, whereas genetic organisation was prepared using snapgene-viewer. ARG, antimicrobial resistance gene; ID, BLAST identity; IME, integrative and mobilisable element; T4SS, type IV secretion system.