Detection of presumed genes encoding beta-lactamases by sequence based screening of metagenomes derived from Antarctic microbial mats

Abstract

Analysis of environmental samples for bacterial antibiotic resistance genes may have different objectives and analysis strategies. In some cases, the purpose was to study diversity and evolution of genes that could be grouped within a mechanism of antibiotic resistance. Different protocols have been designed for detection and confirmation that a functional gene was found. In this study, we present a sequence-based screening of candidate genes encoding beta-lactamases in 14 metagenomes of Antarctic microbial mats. The samples were obtained from different sites, representing diverse biogeographic regions of maritime and continental Antarctica. A protocol was designed based on generation of Hidden Markov Models from the four beta-lactamase classes by Ambler classification, using sequences from the Comprehensive Antibiotic Resistance Database (CARD). The models were used as queries for metagenome analysis and recovered contigs were subsequently annotated using RAST. According to our analysis, 14 metagenomes analyzed contain A, B and C beta-lactamase genes. Class D genes, however, were identified in 11 metagenomes. The most abundant was class C (46.8%), followed by classes B (35.5%), A (14.2%) and D (3.5%). A considerable number of sequences formed clusters which included, in some cases, contigs from different metagenomes. These assemblies are clearly separated from reference clusters, previously identified using CARD beta-lactamase sequences. While bacterial antibiotic resistance is a major challenge of public health worldwide, our results suggest that environmental diversity of beta-lactamase genes is higher than that currently reported, although this should be complemented with gene function analysis. [Figure not available: see fulltext.].