Elucidating environmental resistomes using in silico analysis

Abstract

The antibiotic resistance crisis is one of the most significant challenges in healthcare. There is some evidence suggesting that clinical and environmental resistance may be linked via the horizontal acquisition of resistance genes among pathogens derived from the environment. In addition, anthropogenic influence may contribute to resistance in natural environments. Understanding the nexus of interactions among microbial communities including pathogens, their environments and the concomitant increase in antimicrobial resistance genes is a major ecological endeavour. Yet, despite its importance, resistance is understudied in the environment, and there is a clear knowledge deficit regarding the taxonomic distribution and function of these genes in natural environments. To reduce this knowledge deficit, this study aimed to elucidate environmental resistomes using in silico approaches. We predicted antibiotic resistance genes (ARGs) in prokaryotic genomes acquired from the proGenomes database. These genomes were assigned to eight discrete environments. The taxa IDs from the proGenomes database were assigned taxonomic ranks. Genomic islands harbouring ARGs, and phages were also predicted. We also determined the phylogenetic diversity and the extent to which ARGs, and ARG-carrying taxa were overrepresented in some environments. The analysis showed that host associated environments had the highest richness of ARGs, with soils demonstrating the highest diversity levels. Multidrug efflux pumps were the largest resistance group across all habitats, followed by beta-lactamases. The antibiotic classes, to which the highest resistance was conferred, included penams, cephalosporins, tetracyclines, macrolides, aminoglycosides and fluoroquinolones. The analysis suggests that resistance to tetracycline was the most variable, being high in disease and host associated environments, but lower in other categories. All environments had at least one ARG in a large proportion of their species. Proteobacteria has the most ARGs, with the largest proportion found in Gammaproteobacteria. In total, 116 ARGs were shared, but most large groups and resistance to most antibiotics and classes were shared across all habitats. This result suggests that resistance “functions” are present in all environments, although the specific ARGs conferring them were not necessarily shared. ARGs were also found in the predicted genomic islands and bacteriophages. The ARGs in genomic islands appear to be overrepresented in disease and host associated environments, while ARGs in phages are overrepresented in food and host associated, respectively. We found several genes encoding beta-lactamases and aminoglycoside-modifying enzymes in both genomic islands and bacteriophages. Elfamycin resistance appears to be very common in genomic islands, and in general, although this needs to be experimentally confirmed. The results from this study corroborate previous studies albeit with some caveats linked to the difficulties of standardisation and comparisons. Together, the data suggest that resistance is extensive in the environment, with most functions shared but some differences in the relative abundances. Some ARGs were also present in putative genomic islands and phages, suggesting an increased risk for transfer. Despite methodological differences, the results of this study were consistent with previous findings showing common features of the environmental resistome.