Metagenomic characterization of bacterial community and antibiotic resistance genes in representative ready-to-eat food in southern China

Yiming Li, Weiwei Cao, Shuli Liang, Shinji Yamasaki, Xun Chen, Lei Shi, Lei YeScientific 2020 · DOI

TL;DR

This study uses metagenomics to comprehensively characterize antibiotic resistance genes (ARGs) and bacterial communities in ready-to-eat (RTE) foods from southern China. They found that multidrug resistance genes are the most prevalent, and that the interplay between bacterial communities and mobile genetic elements significantly shapes the resistome in these foods.

What Problem Does This Paper Address?

Antibiotic resistance is a growing global health threat, and food, particularly ready-to-eat (RTE) varieties, is increasingly recognized as a potential reservoir and transmission route for antibiotic-resistant bacteria and their genes. While traditional methods have identified some resistance in food, they are limited in scope and depth. This paper addresses the need for a more comprehensive understanding of the complex microbial landscape and the full spectrum of antibiotic resistance genes present in RTE foods, and how these factors interact. Specifically, it aims to move beyond simply cataloging ARGs to understanding the ecological drivers behind their presence and distribution in these commonly consumed food items.

Methods

The researchers employed a metagenomic sequencing approach, specifically using Illumina Novaseq, to analyze DNA extracted from 18 RTE food samples (meat, vegetables, and fruit). This allowed them to simultaneously characterize both the bacterial community composition (using 16S rRNA gene sequencing) and the antibiotic resistance gene (ARG) profiles.

Key methodological steps included:

  • Quality Control and Host Filtering: Raw sequencing reads were meticulously cleaned to remove low-quality bases and adapter contamination, and host DNA was filtered out.
  • Assembly and Gene Prediction: Cleaned reads were assembled into larger DNA fragments (scaffolds), and open reading frames (ORFs) were predicted to identify potential genes.
  • Bacterial Community Profiling: 16S rRNA gene sequences were compared against the SILVA database for taxonomic classification at phylum, family, and genus levels.
  • ARG Profiling: Predicted ORFs were searched against the Comprehensive Antibiotic Resistance Database (CARD) to identify ARGs. The abundance of ARGs was quantified relative to 16S rRNA gene copies.
  • Mobile Genetic Element (MGE) Analysis: Reads were also searched against databases for integrons, plasmids, and transposons to identify MGEs.
  • Statistical and Network Analyses:
    • Canonical Correspondence Analysis (CCA): Used to explore relationships between bacterial genera, MGEs, and ARG profiles.
  • Variation Partitioning Analysis (VPA): Quantified the relative contributions of bacterial community structure and MGEs to ARG alteration.
  • Spearman's Rank Correlation and Network Analysis: Investigated co-occurrence patterns between specific ARG subtypes and microbial taxa to infer potential host-pathogen relationships.

This metagenomic approach is a significant advancement over traditional culture-based or PCR methods, offering a broader, unbiased view of the microbial and genetic landscape.

Datasets

The study utilized 18 ready-to-eat (RTE) food samples collected from six local supermarkets in southern China. These samples comprised:

  • 8 RTE meat samples (roast pork, roast duck, roast chicken, and salmon sushi)
  • 7 RTE vegetable samples (cucumber, kelp, lotus root, etc.)
  • 3 RTE fruit samples (fresh-cut hami melon)

All sequencing data has been deposited in the European Nucleotide Archive (ENA) under accession code PRJEB33440.

Key Insights and Ideas

  • RTE Foods as ARG Hotspots: The study confirms that RTE foods are significant reservoirs of ARGs, with a wide array of resistance gene subtypes detected.
  • Multidrug Resistance Dominance: Multidrug resistance genes were the most prevalent ARG type across all tested RTE foods, highlighting a broad resistance capacity within these microbial communities.
  • Phylum-Level Consistency, Genus-Level Diversity: While the dominant bacterial phyla (Proteobacteria, Firmicutes, Cyanobacteria, Bacteroidetes, Actinobacteria) were consistent across RTE food types, the family and genus-level compositions showed considerable variation, indicating that food type influences microbial structure at finer taxonomic levels.
  • Proteobacteria as ARG Hosts: ARGs were disproportionately enriched in Proteobacteria compared to other bacterial phyla, suggesting this phylum may play a key role in carrying and disseminating resistance genes.
  • Interplay of Community and MGEs Drives Resistome: The most striking finding is that the combined influence of bacterial community structure and mobile genetic elements (MGEs) is the primary driver of ARG alteration in RTE foods, explaining nearly 80% of the observed changes. This underscores the importance of considering the ecological context and mobile elements when studying antibiotic resistance.
  • Potential Host-Pathogen Links Identified: Network analysis revealed strong correlations between specific ARGs and microbial genera, suggesting potential host organisms for these resistance genes. Genera like Klebsiella, Enterobacter, Pseudomonas, and Acinetobacter were identified as potential hosts for various ARGs, including multidrug resistance genes.

Key Results

  • ARG Diversity: A total of 204 ARG subtypes belonging to 18 ARG types were detected.
  • ARG Abundance: ARG abundance ranged from 2.81 × 10<sup>-5</sup> to 7.7 × 10<sup>-1</sup> copy of ARG per copy of 16S rRNA gene.
  • Dominant ARG Types: Chloramphenicol, macrolide-lincosamide-streptogramin (MLS), multidrug resistance, aminoglycoside, bacitracin, tetracycline, and β-lactam resistance genes were dominant.
  • Multidrug Resistance Gene Abundance: Multidrug resistance genes accounted for a significant proportion of ARGs, with ratios ranging from 3.7 × 10<sup>-2</sup> to 5.8 × 10<sup>-1</sup>.
  • ARG vs. Total Microbial Genes: ARGs constituted 0.86% of total microbial genes in RTE meat, 0.30% in RTE vegetables, and 0.45% in RTE fruit.
  • MGE Abundance Variation: The relative abundance of MGEs varied significantly among RTE food samples.
  • Shared ARGs: 82 ARG subtypes were shared across all RTE food types, with multidrug resistance, bacitracin, aminoglycosides, and tetracycline being the most abundant among these shared genes.
  • Correlation Strength: Strong Spearman's rank correlations (p > 0.8, P < 0.01) were observed between microbial diversity and ARG diversity.
  • VPA Contributions: The joint effect of bacterial community and MGEs contributed 79.37% to resistome alteration, while bacterial community shift alone contributed 18.20%, and MGEs change alone contributed 1.25%.

Limitations and Future Work

  • Sample Size and Geographic Scope: While 18 samples provide a valuable snapshot, the authors suggest that future studies with larger sample sizes and broader geographic coverage are needed to validate these findings and reveal more comprehensive information and biological significance.
  • Food Processing Influence: The study acknowledges that food processing methods can influence microbial communities and ARG profiles, and this factor was not deeply explored.
  • Host Identification Nuances: While network analysis suggests potential ARG hosts, further experimental validation is required to confirm these associations definitively.
  • Tracking Contamination Pathways: The authors emphasize the need for further systematic studies to track contamination pathways of AR bacteria and ARGs in food production and to assess their potential harm to humans.
  • Understanding Bacterial Community's Role: While the study highlights the significant role of bacterial community shifts in ARG alteration, the precise mechanisms by which these shifts drive ARG changes require further investigation.

Summary updated

Abstract

Ready-to-eat (RTE) foods have been considered to be reservoirs of antibiotic resistance bacteria, which constitute direct threat to human health, but the potential microbiological risks of RTE foods remain largely unexplored. In this study, the metagenomic approach was employed to characterize the comprehensive profiles of bacterial community and antibiotic resistance gene (ARG) in 18 RTE food samples (8 RTE meat, 7 RTE vegetables and 3 RTE fruit) in southern China. In total, the most abundant phyla in RTE foods were Proteobacteria, Firmicutes, Cyanobacteria, Bacteroidetes and Actinobacteria. 204 ARG subtypes belonging to 18 ARG types were detected with an abundance range between 2.81 × 10<sup>-5</sup> and 7.7 × 10<sup>-1</sup> copy of ARG per copy of 16S rRNA gene. Multidrug-resistant genes were the most predominant ARG type in the RTE foods. Chloramphenicol, macrolide-lincosamide-streptogramin, multidrug resistance, aminoglycoside, bacitracin, tetracycline and β-lactam resistance genes were dominant, which were also associated with antibiotics used extensively in human medicine or veterinary medicine/promoters. Variation partitioning analysis indicated that the join effect of bacterial community and mobile genetic elements (MGEs) played an important role in the resistome alteration. This study further deepens the comprehensive understanding of antibiotic resistome and the correlations among the antibiotic resistome, microbiota, and MGEs in the RTE foods.
BibTeX
@inproceedings{li2020metagenomic,
  author = {Yiming Li and Weiwei Cao and Shuli Liang and Shinji Yamasaki and Xun Chen and Lei Shi and Lei Ye},
  title = {{Metagenomic characterization of bacterial community and antibiotic resistance genes in representative ready-to-eat food in southern China}},
  booktitle = {Scientific},
  year = {2020},
  doi = {10.1038/s41598-020-72620-4},
}
Figure