Genomic characterization of antimicrobial resistance and mobile genetic elements in swine gut bacteria isolated from a Canadian research farm

Abstract

Introduction: The widespread use of antimicrobials in the livestock industry has raised global concerns regarding the emergence and spread of antimicrobial resistance genes (ARGs). Comprehensive databases of ARGs specific to different farm animal species can greatly improve the surveillance of ARGs within the agri-food sector and beyond. In particular, defining the association of ARGs with mobile genetic elements (MGEs)-the primary agents responsible for the spread and acquisition of resistant phenotypes among bacterial populations-could help assess the transmissibility potential of clinically relevant ARGs. Recognizing the gut microbiota as a vast reservoir of ARGs, we aimed to generate a representative isolate collection and genome database of the swine gut microbiome, enabling high-resolution characterization of ARGs in relation to bacterial host range and their association with MGEs.

Results: We generated a biobank of bacteria from different sections of the gastrointestinal tracts of four clinically healthy pigs housed at a research farm in Ontario, Canada. The culturing was performed under anaerobic conditions using both selective and general enrichment media to ensure the capture of a diverse range of bacterial families within the swine gut microbiota. We sequenced the genomes of 129 unique isolates encompassing 44 genera and 25 distinct families of the swine gut microbiome. Approximately 85.3% (110 isolates) contained one or more ARGs, with a total of 246 ARGs identified across 38 resistance gene families. Tetracycline and macrolide resistance genes were the most prevalent across different lineages of the swine gut microbiota. Additionally, we observed a wide range of MGEs, including integrative conjugative elements, plasmids, and phages, frequently associated with ARGs, indicating that the swine gut ecosystem is conducive to the horizontal transfer of ARGs. High-throughput alignment of the identified ARG-MGE complexes to large-scale metagenomics datasets of the swine gut microbiome suggests the presence of highly prevalent and conserved resistome sequences across diverse pig populations.

Conclusion: Our findings reveal a highly diverse and relatively conserved reservoir of ARGs and MGEs within the gut microbiome of pigs. A deeper understanding of the microbial host range and potential transmissibility of prevalent ARGs in the swine microbiome can inform development of targeted antimicrobial resistance surveillance and disease control programs.

Keywords: Antimicrobial resistance; Integrative and conjugative elements; Mobile genetic elements; Swine gut microbiome.

Fig. 1

Phylogenetic diversity of swine isolate collection. Whole genome sequencing was performed on 129 unique isolates recovered from four healthy pigs. Phylogenetic relatedness among different species was measured by constructing a maximum-likelihood tree using 120 bacterial marker genes of the GTDB-Tk and visualized using the software (https://itol.embl.de). Microbial clades on the tree are coloured according to the taxonomic classification at the family level. The outer rings from inside to outside represent host animal, sample source, GC content, and genome size of each isolate

Fig. 2

Distribution of ARGs across microbial species. The ARGs were identified using the Resistance Gene Identifier (RGI) from the CARD database and validated by NCBI AMRFinderPlus. The ARGs were categorized based on the mechanisms of resistance and the specific antimicrobials to which they confer resistance. Bar charts show the total number of ARGs detected in each species. Microbial clades on the tree are coloured according to the taxonomic classification at the family level

Fig. 3

Classification of ARG-associated MGEs: The tree was created by assessing Mash distances between the flanking regions of each ARG (40 kbp upstream and downstream) using MashTree (https://github.com/lskatz/mashtree). The color of each node corresponds to the phylogenetic lineage of the host species. Signature mobility genes were identified using profile hidden Markov models (pHMM) of relaxase (PF03432), resolvase (PF00239), recombinase (PF07508), phage-like integrase (PF00589), and the type IV secretion system (T4SS obtained from CONJscan). The flanking regions were categorized as Integrated Conjugative Elements (ICEs), Putative ICEs, Integrative Mobile Elements (IMEs), and other putative conjugative MGEs based on the profile of mobility genes. Virsorter2 (v2.2.3), PHASTER, and PLSDB (v2020.11.19) were utilized to detect potential bacteriophage and plasmid-related sequences in the flanking regions of ARGs. The bar chart illustrates the length of sequences obtained from the flanking regions of ARGs, with regions shorter than 80 kbp indicating overlap with contig ends

Fig. 4

Distribution of the identified ARG-associated MGEs in public swine metagenomic datasets. Metagenomics reads from the fecal microbiome of 287 pigs across three countries (France (n = 100), Denmark (n = 100) and China (n = 87)) were mapped to the retrieved fraction of the contigs containing ARGs and signature genes of mobility machinery. Samples were considered to contain the target ARG-associated MGEs if their metagenomes had “good alignment” (defined by greater than 80% coverage of each ARG-MGE sequence or more than 20,000 bp coverage of the extracted ARG-MGE complexes) and “good coverage” (defined by a minimum of 10 × average base coverage across the aligned region). The bar chart on the left indicates the percentage of swine metagenomes with good alignment and coverage (dark green), good alignment but low coverage (light green), good coverage but short alignment region (yellow), and those with poor coverage and alignment (light grey). The bar chart on the right shows the sequence length of the ARG-associate MGEs identified in the present study, colour coded according to MGE classifications