Landscape of mobile genetic elements and their antibiotic resistance cargo in prokaryotic genomes

Abstract

Prokaryotic Mobile Genetic Elements (MGEs) such as transposons, integrons, phages and plasmids, play important roles in prokaryotic evolution and in the dispersal of cargo functions like antibiotic resistance. However, each of these MGE types is usually annotated and analysed individually, hampering a global understanding of phylogenetic and environmental patterns of MGE dispersal. We thus developed a computational framework that captures diverse MGE types, their cargos and MGE-mediated horizontal transfer events, using recombinases as ubiquitous MGE marker genes and pangenome information for MGE boundary estimation. Applied to ∼84k genomes with habitat annotation, we mapped 2.8 million MGE-specific recombinases to six operational MGE types, which together contain on average 13% of all the genes in a genome. Transposable elements (TEs) dominated across all taxa (∼1.7 million occurrences), outnumbering phages and phage-like elements (<0.4 million). We recorded numerous MGE-mediated horizontal transfer events across diverse phyla and habitats involving all MGE types, disentangled and quantified the extent of hitchhiking of TEs (17%) and integrons (63%) with other MGE categories, and established TEs as dominant carriers of antibiotic resistance genes. We integrated all these findings into a resource (proMGE.embl.de), which should facilitate future studies on the large mobile part of genomes and its horizontal dispersal.

Figure 1.

Prokaryotic MGE identification workflow and cargo analysis. (A) The five major recombinase families were annotated in proGenomes2 (22) based on subfamily-level HMMs, validated based on EggNOG (17) and Refseq (55) annotations, as well as by the presence of known associated domains and catalytic site residues. (B) MGE assignment of identified recombinases into six major MGE categories and distinction from cellular recombinases using i) pangenome-based boundary estimation and neighbourhood analysis of core genes (black) and accessory genes (yellow) in 76k genomes, ii) association of MGE types with recombinase subfamilies and iii) rules requiring the presence of MGE type-specific accessory genes. (C) Selected examples of identified MGEs carrying ARG (Antibiotic Resistance Gene) cargo, illustrating different aspects of novelty revealed by in-depth analysis. Catalytic and non-catalytic domains: presence of two recombinase domains one of which (‘rve_3’, crossed) putatively lacks the active site implies functional diversification. Nested MGEs: co-localisation of multiple recombinases belonging to diverse families on recombinase islands. Horizontal Gene Transfer (HGT) between domains of life: the recombinase and almost all neighbouring genes within the MGE show high sequence similarity (> 95% nucleotide sequence identity) between the bacterium Bacteroides ovatusand the archaeon Methanosarcina mazei, implying HGT between domains of life. Novel domain associations: a ‘Streptin-Immun’ domain within a relaxase, confers the host with antibiotic auto-immunity (53).

Figure 2.

A comprehensive prokaryotic MGE census (A) Contribution of each of the five major recombinase families to the six different MGE categories as well as to cellular recombinases. (B) Number of MGEs per category (barplot) and percentage of categories (donut chart) in 76k genomes from 3k species using the workflow described in Figure 1. (C) Predicted lengths of non-nested MGEs in base pairs. The whiskers span from the 10th to the 90th percentile

Figure 3.

Taxonomic distribution of MGEs. (A) Prevalence of MGE categories and dominance of transposable elements across taxonomic classes (with at least 10 genomes), sorted by taxonomic marker gene-based phylogeny (38). (B) Association of MGE categories (average MGE counts per species) with different taxonomic classes (Wilcoxon rank-sum test, * indicates P-value < 0.05 after Bonferroni correction).

Figure 4.

MGE-mediated Horizontal Gene Transfer (HGT). (A) Occurrences of recent MGE-mediated HGT events decrease considerably with taxonomic distance; (B) Overview of MGE-mediated HGT events across a phylogenetic tree (based on phylogenetic marker genes) of taxonomic classes. The heatmap quantifies the HGT events (coloured according to the legend at the bottom) between taxonomic classes and arcs indicate the contribution of transposable elements (in red) over other MGE categories (in grey). (C) Nesting analysis of integrons with different MGE categories shows their significant nesting with CE- Conjugative elements (Fisher's odds ratio 4.3, P-value = 2.9e−86) in HGT subset compared to all nested occurrences. (D) Heatmaps illustrating the promiscuous horizontal transfer of transposable elements across habitats compared to other MGE categories. All MGE categories show high within habitat MGE dynamics (diagonals) compared to between habitat.

Figure 5.

Antibiotic Resistance Genes (ARGs) carrying potential of MGEs. (A) Analysis of MGEs (per category) that carry ARGs show enrichment of ARG in transposable elements compared with other genomics regions, indicating transposable elements as major MGE associated ARG carriers. (B) Association of antibiotic resistance mechanisms with different MGE categories. Asterisks indicate significant enrichment according to one sided Fisher's exact test (P-value < 0.05 after Bonferroni correction).