Novel Soil-Derived Beta-Lactam, Chloramphenicol, Fosfomycin and Trimethoprim Resistance Genes Revealed by Functional Metagenomics

Abstract

Antibiotic resistance genes (ARGs) in soil are considered to represent one of the largest environmental resistomes on our planet. As these genes can potentially be disseminated among microorganisms via horizontal gene transfer (HGT) and in some cases are acquired by clinical pathogens, knowledge about their diversity, mobility and encoded resistance spectra gained increasing public attention. This knowledge offers opportunities with respect to improved risk prediction and development of strategies to tackle antibiotic resistance, and might help to direct the design of novel antibiotics, before further resistances reach hospital settings or the animal sector. Here, metagenomic libraries, which comprise genes of cultivated microorganisms, but, importantly, also those carried by the uncultured microbial majority, were screened for novel ARGs from forest and grassland soils. We detected three new beta-lactam, a so far unknown chloramphenicol, a novel fosfomycin, as well as three previously undiscovered trimethoprim resistance genes. These ARGs were derived from phylogenetically diverse soil bacteria and predicted to encode antibiotic inactivation, antibiotic efflux, or alternative variants of target enzymes. Moreover, deduced gene products show a minimum identity of ~21% to reference database entries and confer high-level resistance. This highlights the vast potential of functional metagenomics for the discovery of novel ARGs from soil ecosystems.

ResFinderFG v2.0: a database of antibiotic resistance genes obtained by functional metagenomics

“…Yet, ResFinderFG v2.0 allowed the detection of a significant proportion of beta-lactamases encoding genes which were not detected with other databases. It was expected since many publications using functional metagenomics reported beta-lactamase encoding genes distant from the ones described in ARG databases 11,14,31,34,46,54,59 and a distant one has been evidenced recently from soil samples 39 . Other antibiotic families were even more specifically associated with ResFinderFG v2.0, such as sulfonamides/trimethoprim, phenicols, glycopeptides/cycloserine resistance genes.…”

Novel amikacin resistance genes identified from human gut microbiota by functional metagenomics

“…Functional metagenomics, which does not rely on known sequences, is recognized as a solid and efficient approach to identifying novel ARGs in the environment and the human microbiota (Fouhy, Ogilvie, et al, 2014a; Fouhy, Ross, et al, 2014b; Marathe et al, 2018; Willms et al, 2021). We constructed a human gut metagenomic library, and five unique amikacin resistance genes were identified, including two novel genes aac(6′)‐ Iao and rmt I. AAC(6′)‐Iao shared 48% identity to AAC(6′)‐Ian identified previously in a clinical isolate (Jin et al, 2015), while RmtI harboured the highest identity of 32% with ArmA in Klebsiella pneumoniae (Galimand et al, 2003).…”

“…Functional metagenomics, which does not rely on known sequences, is recognized as a solid and efficient approach to identifying novel ARGs in the environment and the human microbiota (Fouhy, Ogilvie, et al, 2014a;Fouhy, Ross, et al, 2014b;Marathe et al, 2018;Willms et al, 2021). We constructed a human gut metagenomic library, and five unique amikacin resistance genes were identified, including two novel genes aac(6′)-Iao and rmtI.…”

Ultra-Accurate Classification and Discovery of Functional Protein-Coding Genes from Microbiomes Using FunGeneTyper: An Expandable Deep Learning-Based Framework

“…To validate the capacity of FunGeneTyper models in discovering new functional genes, experimentally confirmed ARGs from functional metagenomics studies were retrieved from NCBI's protein database (accession numbers in Dataset S7). After removing those ARGs showing perfect sequence match to the CARD database (or core dataset of ARGs), 297 experimentally confirmed ARG sequences of human gut 34 (n = 168), WWTPs 35 (n = 77), and soil [36][37][38][39] where TP is the number of true positives, TN is the number of true negatives, FP is the number of false positives, and FN is the number of false negatives.…”

“…Table 1 Comparative performance between FunGeneTyper and other alternative classification tools for the discovery of experimentally confirmed new ARGs. In total, 297 experimentally confirmed ARGs sequences of human gut 34 (n = 168), WWTPs 35 (n = 77), and soil [36][37][38][39] (n = 52) bacteria were included in the analysis which was performed based on default parameters of each tools recommended by the developers.…”