What if the biggest obstacle to tracking antibiotic resistance is not finding bacterial DNA, but assembling the mobile DNA circles that carry it? This study used long-read sequencing to uncover plasmids—and resistance structures—that short reads can miss.

Plasmids are important and versatile bacterial extrachromosomal DNA molecules, and they have been a central topic in bacterial genetics and biology. However, short-read high-throughput sequencing methods cannot reliably assemble plasmids, making it difficult to investigate the diversity of plasmid structures and functions.

Besides the main chromosome, bacteria often carry extrachromosomal circular DNA molecules, the plasmids, for additional functions that are not encoded on the chromosome. Plasmids are the variable parts of the bacterial genome, bringing quick adaptation, high versatility, and beneficial traits for combating a changing environment with a small genome.

Plasmids can be replicated and transferred between bacterial cells, leading to very quick acquisition of functions. Acquisition of antibiotic resistance genes hosted on plasmids is a primary route for antimicrobial resistance by bacteria. The short reads of second-generation sequencing methods make it very difficult to detect, let alone assemble, plasmids.

As a result, knowledge of plasmid sequence variability and the diversity of plasmid functions has been limited. Long-read third-generation sequencing methods can produce reads of up to several megabasepairs and can sometimes sequence a whole plasmid, normally less than two hundred kilobases in size, in one read.

This work aims to understand how much plasmid diversity and function remain poorly understood, and how improved third-generation sequencing technologies can address the issue. The approach was to determine plasmids from bacterial strains in isolated clinical samples, check how many had not been previously observed, and identify new plasmid structures and functions.

Samples from perianal abscess-suffering patients were used because the disease is caused by bacterial infection and commonly applied antibiotic therapy might result in stress leading to enriched plasmid occurrence. Samples from eleven patients with perianal abscess were collected from March to November, twenty twenty-three, at Qilu Hospital in Qingdao of Shandong University.

For each patient, a sterile swab sampled anal skin, feces, and abscess, after which the samples were stored in sterile ten percent glycerol solution and transported to the laboratory at four degrees Celsius. Dorado performed basecalling and sequence demultiplexing, while NanoQC and Chopper evaluated and filtered sequence quality.

Flye and Medaka assembled and corrected whole genome sequences, and Flye also determined sequence circularity. Busco, Checkm2, and Quast evaluated the assembled sequences, while GTDB-Tk determined taxonomy and AMRFinder identified antibiotic resistance genes.

PlasmidFinder typed plasmids, Prokka performed genome annotation, and Blast searched putative plasmid sequences against the NT database. With third-generation Nanopore sequencing, high-quality whole-genome sequences were obtained for all the isolated bacterial strains, with an average of three point one five contigs assembled per bacterium.

Perfect genome assemblies, in which all DNA molecules were assembled to circular form, were obtained for twenty of the thirty-three bacteria. Of all one hundred four assembled contigs, seventy-six, or seventy-three point one percent, were circular.

These statistics confirm the high quality of the genome sequences and assemblies. Figure one groups plasmids into four size ranges, separately showing all plasmids in panel A and unique new plasmids in panel B. In both panels, the ten-to-one-hundred-kilobase category contains the most plasmids, while the smallest category contains only a few; the authors report an average size of eighty point eighty-nine kilobases for all plasmids and fifty-eight point eighty kilobases for the new ones.

This distribution highlights the prevalence and size diversity of plasmids revealed by long-read sequencing. On average, each strain contains one point six eight plasmids, and most plasmids are over ten kilobasepairs, with an average size of eighty point eighty-nine kilobases.

Very small plasmids, less than one kilobasepair, also exist, with three counts. Their existence was determined by circularity of sequenced DNA molecules, the presence of protein-coding genes, and generally high copy numbers.

However, circular DNA can also indicate excised products made by recombinases in vivo, so the prediction of these very small plasmids may not be accurate. Twelve small contigs in five strains were assembled as parts of a plasmid, and more plasmids may still not have been fully assembled.

A total of twelve new plasmids were found, and the distribution of their sizes was close to that of all plasmids found, with an average size of fifty-eight point eighty kilobases. All twelve new plasmids were assembled to circularity, confirming the quality of the sequences.

The new plasmids included four high-copy plasmids, with predicted copy numbers greater than ten, and eight low-copy plasmids, with predicted copy numbers less than ten. Figure five compares the circular structures of five plasmids with their closest homologs, using colored arrows to mark coding regions, mobility-related genes in green, and replication genes in red.

The comparisons reveal very different organizational patterns: pN17631 is a small three-thousand-one-hundred-thirty-three-base-pair plasmid, while pN19332 spans seventy-nine-thousand-two-hundred-forty-one base pairs, and pN18311 shows extensive rearrangement relative to pYK5. The authors use these maps to illustrate plasmid diversity, shared backbones, and recombination that can shape mobility and function.

The three newly identified large plasmids, larger than one hundred kilobases, are mobile multidrug-resistant plasmids that encode type four secretion systems. All of them encode multiple Rep proteins, suggesting that multiple recombination or integration events led to these complex plasmids.

All three plasmids are IncF plasmids. Figure seven compares the gene neighborhoods carrying mcr-ten-point-one on two plasmids: pN5912 from Enterobacter kobei and pN5921 from Citrobacter portucalensis. In both clusters, xerC sits next to mcr-ten-point-one, while the surrounding genes differ, including insertion-sequence and transposase-associated elements such as IS26, IS4321, and Tn3.

This matters because it documents polymyxin resistance on plasmids in these organisms and highlights the genetic context that may support its mobility. A new mcr-10.1-harboring mobile plasmid was found in Citrobacter portucalensis N59-2.

This antibiotic resistance gene confers resistance to polymyxin, an antibiotic of last resort that has received substantial attention. Previously reported experimentally verified mcr-10.1-harboring plasmids were hosted by Enterobacter cloacae, Enterobacter roggenkampii, and Klebsiella strains.

No mcr-10.1-hosting Citrobacter portucalensis strains had previously been reported. The new conjugation-apparatus-coding, mcr-10.1-hosting Citrobacter portucalensis pN59-2 plasmid expands knowledge of the spectrum of mcr-10.1 hosts.

AMRFinder was used to investigate the presence of antibiotic resistance genes in the isolated strains. Of all strains sequenced, only one Escherichia coli strain did not carry any antibiotic resistance genes. Most of the antibiotic resistance genes were plasmid-borne, suggesting extensive prevalence of antibiotic resistance genes and antimicrobial resistance.

The study examined near-perfect, third-generation-sequencing-based whole-genome sequences from thirty-three bacterial strains from eleven perianal abscess-suffering patients, with particular focus on plasmid diversity. Twelve new plasmids that had not been observed before were identified, possibly because long-read sequencing enabled high-accuracy discovery and surveillance of plasmids.

Compared with their closest homologs, the new plasmids showed significant differences, further confirming the strong dynamics of plasmids and their role in genomic plasticity. Two polymyxin-resistance-conferring, mcr-10.1-carrying conjugative plasmids were found in Citrobacter portucalensis and Enterobacter kobei strains.

This was the first finding of mcr-10.1 in Citrobacter portucalensis and one of the first few evidences of conjugative, mobile mcr-10.1-carrying plasmids. The work shows the strengths of long-read sequencing for identifying plasmids, as well as the lack of knowledge about plasmids and the diversity of plasmid structures.

Long-read Nanopore sequencing recovered a diverse plasmid landscape from thirty-three bacterial strains, including twelve new plasmids and two mobile mcr-10.1-carrying plasmids. The result is a warning that plasmid diversity remains substantially under-surveilled.