Distribution and quantification of antibiotic resistant genes and bacteria across agricultural and non-agricultural metagenomes

Abstract

There is concern that antibiotic resistance can potentially be transferred from animals to humans through the food chain. The relationship between specific antibiotic resistant bacteria and the genes they carry remains to be described. Few details are known about the ecology of antibiotic resistant genes and bacteria in food production systems, or how antibiotic resistance genes in food animals compare to antibiotic resistance genes in other ecosystems. Here we report the distribution of antibiotic resistant genes in publicly available agricultural and non-agricultural metagenomic samples and identify which bacteria are likely to be carrying those genes. Antibiotic resistance, as coded for in the genes used in this study, is a process that was associated with all natural, agricultural, and human-impacted ecosystems examined, with between 0.7 to 4.4% of all classified genes in each habitat coding for resistance to antibiotic and toxic compounds (RATC). Agricultural, human, and coastal-marine metagenomes have characteristic distributions of antibiotic resistance genes, and different bacteria that carry the genes. There is a larger percentage of the total genome associated with antibiotic resistance in gastrointestinal-associated and agricultural metagenomes compared to marine and Antarctic samples. Since antibiotic resistance genes are a natural part of both human-impacted and pristine habitats, presence of these resistance genes in any specific habitat is therefore not sufficient to indicate or determine impact of anthropogenic antibiotic use. We recommend that baseline studies and control samples be taken in order to determine natural background levels of antibiotic resistant bacteria and/or antibiotic resistance genes when investigating the impacts of veterinary use of antibiotics on human health. We raise questions regarding whether the underlying biology of each type of bacteria contributes to the likelihood of transfer via the food chain.

Figure 1. Metagenome samples used in this study.

All metagenomes are publicly available on the MG-RAST website. Superscript numbers denote citations listed in the references section. MG-RAST accession numbers and metagenome statistics are listed in Table S1. Metagenome locations were determined using the latitude and longitude metadata linked to each metagenome in MG-RAST.

Figure 2. Percent of metagenome coded for by genes responsible for Resistance to antibiotic and toxic compounds.

Percent of RATC genes, based on the total number of classified sequences in each metagenome. Metagenomes from microbially dense habitats, such as feces, have a larger proportion of RATC genes compared to metagenomes from habitats that are less microbially dense.

Figure 3. Principal component analysis of 26 metagenomes based on RATC genes.

Information regarding resistance to antibiotic and toxic compounds (RATC) was used to perform a principal component analysis of 26 publicly available metagenomes. Both animal agriculture and human associated metagenomes group in quadrant three. Eigenvectors for quadrant three included tetracycline, aminoglycoside adenyltransferase, and multi-drug resistance efflux pump genes.

Figure 4. Bacteria responsible for betalactamase genes in 16 metagenomes.

Percent of betalactamase genes within each metagenome that are assigned to the listed taxa. For example, 60% of all betalactamase genes in kimchi are associated with Clostridia, 30% of all betalatamase genes in kimchi are associated with Bacilli, and 10% of all betalactamase genes in kimchi are associated with gammaproteobacteria. The alphaproteobacteria carry a larger proportion of the betalactamase genes in marine metagenomes (between 20–36%), compared to metagenomes from food producing animals (1–15%). In contrast, the Clostrida from food producing animals appear to carry proportionally more betalactamse genes than the Clostridia from marine metagenomes. Color formatting indicates low and high values. All true zero values are in green. Values between 0 and 0.5 are listed as 0, but are formatted in yellow. Red indicates the highest value within the chart.

Figure 5. Bacteria responsible for multi-drug resistant efflux pump (MDR efflux) genes in 16 metagenomes.

Percent of MDR efflux genes within each metagenome that are assigned to the listed taxa. For example, 71% of the MDR efflux pump genes in kimchi are associated with Bacilli, 20% are associated with Clostridia, and the remaining MDR efflux pump genes are spread across five other bacterial classes. The Clostridia are the main carriers of MDR efflux genes in cattle metagenomes, while the Bacteroidetes are the main carries of MDR efflux genes in the chicken metagenome. Color formatting indicates low and high values. All true zero values are in green. Values between 0 and 0.5 are listed as 0, but are formatted in yellow. Red indicates the highest value within the chart.

Figure 6. Bacteria responsible for fluoroquinolone resistance genes in 16 metagenomes.

Percent of fluoroquinolone resistance genes within each metagenome that are assigned to the listed taxa. For example, almost all (98%) of the fluoroquinolone resistance genes in the kimchi metagenome were associated with Bacilli. The fluoroquinolone resistance genes in food animal metagenomes were found mainly in Clostridia and Bacteroidetes, while the fluoroquinolone resistance genes in marine metagenomes were found primarily in Alphaproteobacteria. The two Antarctic Lake metagenomes also had a high proportion of their fluoroquinolone resistance genes assigned to Clostridia, but unlike the food animal metagenomes, there were very few fluoroquinolone resistant Bacteroidetes. Color formatting indicates low and high values. All true zero values are in green. Values between 0 and 0.5 are listed as 0, but are formatted in yellow. Red indicates the highest value within the chart.

Figure 7. Bacteria responsible for tetracycline resistance genes in 16 metagenomes.

Percent of tetracycline resistance genes within each metagenome that are assigned to the listed taxa. For example, 70% of all tetracycline resistance genes in the kimchi metagenome were assigned to Bacilli. Tetracycline genes associated with Bacteroidetes are fairly common in food animal and human associated metagenomes, but were not found in any of the marine or environmental metagenomes examined in this study. A similar pattern was seen for tetracycline resistance genes associated with Epsilonproteobacteria, though the overall percent of tetracycline resistance genes associated with this group was much lower. Color formatting indicates low and high values. All true zero values are in green. Values between 0 and 0.5 are listed as 0, but are formatted in yellow. Red indicates the highest value within the chart.

Figure 8. Bacteria responsible for vancomycin resistance genes in 16 metagenomes.

Percent of vancomycin resistance genes within each metagenome that are assigned to the listed taxa. For example, 73% of the vancomycin resistance genes in kimchi were attributed to Bacilli. As expected, the vancomycin resistance genes are carried predominantly by members of the phylum Firmicutes. Color formatting indicates low and high values. All true zero values are in green. Values between 0 and 0.5 are listed as 0, but are formatted in yellow. Red indicates the highest value within the cart.