Mobile genetic element insertions drive antibiotic resistance across pathogens

Matthew G. Durrant, Michelle M. Li, Ben Siranosian, Ami S. BhattbioRxiv 2019 · DOI

TL;DR

This paper introduces "mustache," a novel computational approach to precisely identify mobile genetic elements (MGEs) and their insertion sites directly from short-read sequencing data, without relying on pre-existing databases. This method reveals that MGE insertions are a significant driver of antibiotic resistance evolution in bacterial pathogens, influencing both laboratory adaptation and clinical outcomes.

What Problem Does This Paper Address?

Detecting mobile genetic elements (MGEs) like insertion sequences (IS) and transposons in bacterial genomes is crucial for understanding adaptation and evolution, especially concerning antibiotic resistance. However, existing methods face challenges: they often require curated databases of known MGEs, or rely on "split-read" alignments that assume the inserted element is already present in the reference genome. This limits their ability to discover novel MGEs and accurately pinpoint their insertion sites, particularly in diverse bacterial populations. The authors aim to overcome these limitations with a high-throughput, unbiased, de novo approach.

Methods

The core of this paper is the development and application of a new computational tool called "mustache." The pipeline works as follows:

  1. Candidate Insertion Site Identification: Short reads are aligned to a reference genome. "Mustache" identifies potential insertion sites by looking for reads with clipped ends at the same location, indicating a break where an element has been inserted.
  2. Flank Sequence Consensus: Consensus sequences are generated from these clipped ends, representing the "flanks" of the inserted element.
  3. Element Inference: The identity of the inserted sequence is inferred through several methods:
  4. Assembly-based: Aligning flanks to the assembled genome of the isolate.
  5. Overlap-based: Finding high-confidence overlaps between the two flank ends.
  6. Reference-based: Aligning flanks to a reference genome.
  7. Database-based: Using a dynamically created database of previously identified MGEs.
  8. Clustering and Genotyping: Inferred elements are clustered based on sequence similarity to collapse redundancy. Insertions are then assigned a genotype based on the confidence of their inference.
  9. Hotspot Analysis: The method identifies "hotspots" – genomic regions with a statistically significant enrichment of MGE insertions.

This de novo approach is flexible, can identify novel MGEs, and is sensitive enough for both lab and environmental samples.

Datasets

The study utilized several large datasets:

  • Publicly Available Bacterial Isolates: 12,419 sequenced isolates from nine prevalent bacterial pathogens were analyzed. These were randomly selected from the Sequence Read Archive (SRA).
  • Adaptive Laboratory Evolution (ALE) Experiments:
    • Data from the "megaplate" experiment by Baym et al. (2016) involving E. coli adapting to trimethoprim.
  • Data from the "morbidostat" experiment by Toprak et al. (2011) involving E. coli adapting to chloramphenicol, doxycycline, and trimethoprim.
  • Clinical E. coli Isolates:
    • A collection of 241 E. coli bacteremia isolates (the "Hospital" collection) from Oxford University Hospitals.
  • A collection of 260 E. coli clinical isolates (the "MDR" collection) from the FDA-CDC Antimicrobial Resistance Bank.

Key Insights and Ideas

  • MGEs are not just "junk DNA": The study strongly demonstrates that MGEs are active players in bacterial evolution, directly contributing to adaptation and the acquisition of crucial traits like antibiotic resistance.
  • Target-site specificity matters: Many MGEs exhibit a preference for inserting into specific genomic locations, often near genes involved in important biological functions, suggesting a non-random, potentially adaptive process.
  • A new tool for discovery: "Mustache" provides a much-needed method for unbiased, high-throughput identification of MGEs and their insertion sites, opening doors to discovering novel elements and understanding their roles.
  • Species-specific MGE activity: The repertoire and activity of MGEs vary significantly between bacterial species, indicating different evolutionary strategies. Some species, like E. faecium, show remarkably high MGE activity.
  • MGEs as drivers of resistance: The study provides compelling evidence, from both lab evolution experiments and clinical isolates, that MGE insertions are a significant mechanism for acquiring antibiotic resistance.

Key Results

  • The "mustache" tool demonstrated high sensitivity (up to 88%) and precision (over 98%) in simulations.
  • Analysis of 12,419 isolates across nine bacterial species identified hundreds of known and novel MGEs.
  • MGE insertion hotspots were frequently found near genes involved in antibiotic resistance mechanisms.
  • In E. coli ALE experiments, MGE insertions accounted for a substantial proportion of adaptive mutations conferring antibiotic resistance (e.g., 17.1% more adaptive mutations in TMP adaptation).
  • A genome-wide association study (GWAS) using MGE insertions as features identified significant associations with antibiotic resistance phenotypes in clinical E. coli isolates, highlighting genes like acrR and ompF.
  • The study found that MGE activity varies significantly across species, with E. faecium showing particularly high activity.

Limitations and Future Work

  • Short-read limitations: The authors acknowledge that inferring full inserted sequences from short reads has inherent limitations in precision, especially for MGEs that are difficult to assemble. Future work with long-read sequencing could provide orthogonal validation.
  • Computational intensity: The de novo assembly step can be computationally intensive, potentially limiting application to organisms with very large genomes.
  • Statistical significance: Some associations found in the GWAS did not meet the most stringent significance cutoffs, and the effects were relatively small compared to other genetic factors.
  • Experimental validation: The authors emphasize that the statistically significant associations between gene disruption and antibiotic resistance require orthogonal testing through molecular experiments (gene knockout and rescue).
  • Future directions: The authors suggest further investigating the identified candidate genes for functional significance and exploring the application of their method to metagenomic datasets to study MGEs in complex microbial communities. They also highlight the potential of sequence-specific transposases for genetic engineering applications.

Summary updated

Abstract

Abstract Mobile genetic elements contribute to bacterial adaptation and evolution; however, detecting these elements in a high-throughput and unbiased manner remains challenging. Here, we demonstrate a de novo approach to identify mobile elements from short-read sequencing data. The method identifies the precise site of mobile element insertion and infers the identity of the inserted sequence. This is an improvement over previous methods that either rely on curated databases of known mobile elements or rely on ‘split-read’ alignments that assume the inserted element exists within the reference genome. We apply our approach to 12,419 sequenced isolates of nine prevalent bacterial pathogens, and we identify hundreds of known and novel mobile genetic elements, including many candidate insertion sequences. We find that the mobile element repertoire and insertion rate vary considerably across species, and that many of the identified mobile elements are biased toward certain target sequences, several of them being highly specific. Mobile element insertion hotspots often cluster near genes involved in mechanisms of antibiotic resistance, and such insertions are associated with antibiotic resistance in laboratory experiments and clinical isolates. Finally, we demonstrate that mutagenesis caused by these mobile elements contributes to antibiotic resistance in a genome-wide association study of mobile element insertions in pathogenic Escherichia coli . In summary, by applying a de novo approach to precisely identify mobile genetic elements and their insertion sites, we thoroughly characterize the mobile element repertoire and insertion spectrum of nine pathogenic bacterial species and find that mobile element insertions play a significant role in the evolution of clinically relevant phenotypes, such as antibiotic resistance.
BibTeX
@misc{durrant2019mobile,
  author = {Matthew G. Durrant and Michelle M. Li and Ben Siranosian and Ami S. Bhatt},
  title = {{Mobile genetic element insertions drive antibiotic resistance across pathogens}},
  year = {2019},
  doi = {10.1101/527788},
  howpublished = {bioRxiv},
}
Figure