https://orcid.org/0000-0001-8485-9143
Figures
Abstract
Objectives
Antimicrobial-resistant bacteria are a major global health threat. Mobile genetic elements (MGEs) have been crucial for spreading resistance to new bacterial species, including human pathogens. Understanding how MGEs promote resistance could be essential for prevention. Here we present an investigation of MGEs and their association with resistance genes in pathogenic bacteria collected from 59 diagnostic units during 2020, representing a snapshot of clinical infections from 35 counties worldwide.
Methods
We analysed 3,095 whole-genome sequenced clinical bacterial isolates from over 100 species to study the relationship between resistance genes and MGEs. The mobiliome of Staphylococcus aureus, Enterococcus faecalis, Escherichia coli, and Klebsiella pneumoniae were further examined for geographic differences, as these species were prevalent in all countries. Genes potentially mobilized by MGEs were identified by finding DNA segments containing MGEs and ARGs preserved in multiple species. Network analysis was used to investigate potential MGE interactions, host range, and transmission pathways.
Results
The prevalence and diversity of MGEs and resistance genes varied among species, with E. coli and S. aureus carrying more diverse elements. MGE composition differed between bacterial lineages, indicating strong vertical inheritance. 102 MGEs associated with resistance were found in multiple species, and four of these elements seemed to be highly transmissible as they were found in different phyla. We identified 21 genomic regions containing resistance genes potentially mobilized by MGEs, highlighting their importance in transmitting genes to clinically significant bacteria.
Conclusion
Resistance genes are spread through various MGEs, including plasmids and transposons. Our findings suggest that multiple factors influence MGE prevalence and their transposability, thereby shaping the MGE population and transmission pathways. Some MGEs have a wider host range, which could make them more important for mobilizing genes. We also identified 103 resistance genes potentially mobilised by MGEs, which could increase their transmissibility to unrelated bacteria.
Citation: Johansson MHK, Petersen TN, Nag S, Lagermann TMR, Birkedahl LE, Tafaj S, et al. (2025) Investigation of mobile genetic elements and their association with antibiotic resistance genes in clinical pathogens worldwide. PLoS One 20(8): e0330304. https://doi.org/10.1371/journal.pone.0330304
Editor: Mabel Kamweli Aworh,, Fayetteville State University, UNITED STATES OF AMERICA
Received: March 14, 2025; Accepted: July 29, 2025; Published: August 18, 2025
Copyright: © 2025 Johansson et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All sequenced genomes are available from the European Nucleotide Archive database (https://www.ebi.ac.uk/ena/browser/view/ERP141886).
Funding: This work was supported by the Novo Nordisk Foundation (Grant: NNF16OC0021856: Global Surveillance of Antimicrobial Resistance) and the European Union’s Horizon 2020 research and innovation program (Grant: 874735). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The emergence of antimicrobial-resistant bacteria (AMR) is recognised as a significant threat to global public health [1,2] and estimates suggest that 1.3 million deaths annually can be attributed to AMR [3].
Several bacteria, such as the ESKAPE [4] group pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumanii, Pseudomonas aeruginosa, and Enterobacter spp) and Escherichia coli have been increasingly involved in resistant infections [5]. These bacteria can be difficult to treat since they frequently carry antimicrobial resistance genes (ARG), often acquired via horizontal gene transfer and mediated by various mobile genetic elements (MGE) [6–9].
MGEs are discrete regions of DNA that can promote their transposition or the transposition of other elements between DNA molecules. They are classified based on their properties and genetic layout into types [6]. Inter-cellular transposing MGEs such as plasmids, integrative and conjugative elements (ICE), and integrative and mobilisable elements (IME) can either conjugate or be mobilised by the conjugation of other elements [7–9]. They often carry other ARGs or intra-cellular transposing MGEs. Unlike plasmids, ICEs and IMEs are primarily integrated into the host chromosome [10]. MGEs integrated into a host’s DNA are referred to as integrated MGEs (iMGEs).
Insertion sequences (IS) are among the smallest types of intra-cellular transposing iMGEs, often consisting of a transposase bounded by inverted repeats (IRs) [11,12]. They are typically unable to carry accessory genes but can facilitate gene mobility by forming composite transposons (comTn) structures [13, 14]. Unit transposons (Tn) are a diverse type of iMGEs whose members are frequently associated with ARGs. Several members carry integrons [15,16], a type of iMGE that can rapidly exchange their carried accessory genes [17]. Extensive literature describes the different types of MGEs [6,9] and their association with AMR in various species [18–20].
The interplay of different MGEs forms a complex transposition network that has been essential for recruiting ARGs [7, 21] and spreading them to infectious bacteria [22,23]. MGEs are also thought to help retain resistance genes in environments with low levels of antimicrobials [24].
Here, we present an investigation of resistance determinants and their association with MGE from 59 diagnostic units worldwide. Samples were collected in 2020 from patients with symptoms of disease caused by a pathogen, and thus, whole genome sequencing was used for diagnostic analysis. These samples provide a valuable insight into the association between ARGs and MGEs as they represent an unbiased snapshot of clinically relevant bacteria isolated in 2020. The data encompasses a diverse set of bacterial species, lineages, and tissue types and should therefore better approximate the diversity of resistance determinants and mobilome.
Methods
Dataset selection
Clinical pathogens were collected as part of the Two Weeks in the World (TWIW) research project by participating diagnostic units during 2020 and sent to the Group for Genomic Epidemiology, Technical University of Denmark (DTU), for whole genome sequencing. Protocols for DNA sampling are available at the TWIW website (https://twiw.genomicepidemiology.org/about), and the sequences are available at ENA (see S1 Appendix for sample information and accession numbers). The TWIW dataset comprises 3,095 sequenced isolates, representing 115 species from 35 countries [25].
We primarily investigated E. coli, K. pneumoniae, E. faecium, and S. aureus isolates as they were the most prevalent Gram- and Gram+ species in this dataset with a good geographic representation. The sequences were processed, quality-controlled, and de-novo assembled using the method outlined in the supplementary methods section. A total of 1,191 samples (553 E. coli, 227 K. pneumoniae, 77 E. faecium, 334 S. aureus), representing 35 countries from six WHO-defined regions [23] and is now referred to as the “primary” dataset.
The remaining sequences from less frequent species were used to verify putative transposable elements and to investigate the phylogenetic confinement of iMGEs. These sequences were processed using the previously described method, creating a “secondary dataset” consisting of 1,559 samples from 101 species.
Read processing, genome assembly, and QC filtering
Only isolates sequenced using the Illumina NextSeq 500 platform were included to mitigate bias introduced by differences in sequencing chemistry.
Reads were trimmed using bbduk2 (part of BBmap version 36.49) (score cutoff 20 and removing reads shorter than 50 bp), and adapter sequences were removed by matching using an internal adapter database [26]. Read quality was assessed with FastQC [27] (version 0.11.5) before and after read processing. The processed reads were de-novo assembled with Spades [28] version 3.11.0 using error correction, coverage cutoff = 2, and k-mer sizes 21, 33, 55, 77, 99, and 127. Assembly quality metrics were calculated with Quast [29] version 4.5. If the isolate had been sequenced multiple times (47 isolates), the sequencing run with the highest-quality bases was selected.
Kmerfinder [30] and ribosomal Multi Locus Sequence Typing (rMLST) were used to verify project participants’ species annotations and identify contaminated samples. Isolates in which the predicted species differed from the annotation were excluded, as they could have been mixed up.
Kmerfinder and rMLST were run using the default thresholds using the processed reads as input. A species was said to be verified if Kmerfinder and rMLST predictions were in concordance, and the main species prediction exceeded 80% rMLST support and 40% Kmerfinder coverage. Isolates predicted to contain multiple species (Kmerfinder coverage > 1% or rMLST support > 1%) were excluded since they could be contaminated.
Isolates with an unexpectedly large assembly size were removed, as it could indicate that they contained multiple strains, i.e., if the assembly size was larger than 99.7% (3 std) of the reference genomes for that species in NCBI RefSeq. Genome sizes of non-atypical genomes with the completeness of “chromosome” or “complete assembly” were used (data accessed 2023-03-17).
The assembly completeness was estimated from the proportion of non-identified loci in the core genome Multi Locus Sequence Typing (cgMLST) profiles. chewBBACA [31] version 3.0.0 with the relevant schemas from cgmlst.org [32–34] was used for typing. Isolates with < 95% of the core genome were excluded.
A maximum of 15 isolates per species and city was included to reduce the overrepresentation of some locations. The final dataset consisted of 1,191 isolates from 35 countries covering seven geographical world regions.
Controlling for high relatedness in the dataset
To evaluate the diversity of the dataset and ensure it reflects the strains typically encountered in healthcare settins, we employed a combination of cgMLST and Multi Locus Sequence Typing (MLST) typing. The within-species diversity was estimated using both the cgMLST and MLST profiles. Overrepresented lineages were identified from the MLST sequence types (ST) using the mlst [35,36] tool version 2.23.0. Isolates were clustered based on their cgMLST profile with scipy [23] version 1.10.1 using Jaccard distance and visualised using the toolkit for Python.
Each species was represented by isolates from several MLST sequence types (E. coli 94, K. pneumoniae 95, E. faecalis 31, and S. aureus 56), of which approximately 44%−76% were considered rare as they were found in less than 5% of all isolates (S1 Fig). The high average pairwise cgMLST allele difference for all species also suggested that the dataset is highly diverse.
In-silico prediction of AMRs and iMGEs
ARGs were predicted with Resfinder (tool version 4.3.1, ResfinderDb version 2.1.0, disinfinderDb version 2.0.0) with default settings, using the assembled contigs as input [37]. An updated version of MobileElementFinder [38] (version 1.1.2) was used to predict miniature inverted-repeats transposable elements (MITEs), IS, Tn, ICE, IME, and cis-mobilizable elements (CIMEs). This new version adds 1,686 IS and 70 Tn to its database.
Identification of plasmid-borne iMGEs and ARGs
Contigs were classified as being of either plasmid or chromosomal origin using the consensus prediction of PlasClass [39] version 0.1.1 (threshold 0.5) and Platon [40] version 1.6 (executed in accuracy mode with either the Enterococcus, Staphylococcus, or Bacteria database).
A two-way ANOVA (statsmodels python package) was used to evaluate differences in the amount of plasmid DNA between the species. The normality of residuals and heterogeneity of variance were ensured.
Data visualisation
Unless otherwise stated, all visualisations were created using a combination of Matplotlib [41] version 3.7.1 and Seaborn [42] version 0.12.2 for Python 3.11. Geographical maps were created using GeoPandas version 0.13.0 using shapefiles from Natural Earth.
Investigation of the iMGE population, geography, and lineage
Isolates were clustered using scipy [23] on the iMGE profile (Jaccard distance and average linkage) to investigate differences between geographical regions and lineages.
Identification of iMGEs associated with ARGs and putative mobilised complexes
iMGEs have the potential to mobilize nearby genes by forming comTns. Potentially mobilized ARGs were identified in the assembled 1,191 primary samples by locating iMGEs within 10 kb of an ARG, which corresponds with the largest compTn in the MGEdb database. Combinations of iMGE and ARG were kept if the distance between the elements and the intermediary sequence was preserved in multiple samples. The sequence similarity was calculated using Sourmash [43] version 4.8.2.
Finding the same preserved combination of iMGE and ARG, or putative translocatable unit (pTU), in multiple species could indicate mobilization. The presence of the pTUs on other species was investigated by searching for similar sequences in the secondary dataset using blastn [44] version 2.14.0. The pTU was queried against a custom database of the assembled genomes, and hits with greater than 95% query coverage and greater than 90% sequence identity were considered matches.
Investigation of the iMGE host range
The presence of iMGEs in bacterial species was used to estimate host range and possible transposition pathways. iMGEs were said to be confined to a taxonomic level if they were only found in bacteria of the same taxa. Hypothetical transposition pathways were modelled as a bi-directional multi-graph where species constituted nodes and iMGEs edges using NetworkX [45] version 3.1. Communities of densely connected bacterial species were identified using the Louvain method [46] of community detection.
Results
Here, we present an analysis of mobilised ARGs in 1,191 clinical isolates (primary dataset) from the four common pathogens E. coli (553), K. pneumoniae (227), E. faecalis (77), and S. aureus (334), representing a snapshot of the global diversity of resistance determinants (S1 Appendix).
The data is a subset of a larger dataset containing 3,095 clinical isolates, including 115 species worldwide [47] as seen in S2 Fig. All isolates are available on ENA (accession number: ERP141886).
Species-specific differences in the ARG profile
A total of 7,159 ARGs were identified in the primary dataset. Among the four pathogens, Gram- bacteria exhibited greater ARG diversity, with K. pneumoniae carrying the most diverse set of resistance genes (S3b Fig). Beta-lactam resistance genes were highly prevalent in K. pneumoniae (99.6% of all isolates), followed by S. aureus (79.3%) and E. coli (58.6%). In contrast, no beta-lactam resistance genes were in any of the E. faecalis isolates, and in S. aureus, it was primarily mediated by blaZ and mecA.
The distribution of beta-lactamase genes varied geographically, particularly in E. coli and K. pneumoniae. In E. coli, blaTEM genes were widespread across all continents, whereas blaSHV was more prevalent in central Europe. The blaCTX-M and blaOXA gene families were more prevalent in southern Europe, Turkey, and Africa (Fig 1).
Countries with resistance are coloured in blue. The pie chart shows the composition of gene families per country. Map was created using shapefiles from Natural Earth.
Geographic differences were also evident in the distribution of gene variants. For instance, blaCTX-M-15 and blaCTX-M-27 were found on almost every continent, whereas blaCTX-M-1 was only found in Europe, and blaCTX-M-24 was only identified in isolates from Oceania (S1 Table). The blaCMY genes were only found in isolates from Asian and African countries, with Asian isolates exhibiting higher gene diversity than those observed in Africa. In contrast, African isolates only carried blaCMY-2.
In K. pneumoniae, beta-lactam resistance was primarily mediated by blaSHV, blaTEM, and blaCTX genes. Notably, isolates from South America frequently harboured primarily blaKPC and blaNDM genes (Fig 2 and S2 Table), highlighting regional differences in resistance mechanisms.
Countries with resistance are coloured in blue. The pie chart shows the composition of gene families per country. Map was created using shapefiles from Natural Earth.
Approximately 75% of the E. faecalis isolates were predicted to carry tetracycline resistance genes (mediated by tet(M) and tet(L)), while about 44.1% of the isolates carried aminoglycoside resistance genes. Macrolide (~35.1%), glycopeptides (~2.6%), and amphenicol (~1.2%) resistance genes were only present in a small number of isolates (S4 Fig). The intrinsic Lsa family of efflux pumps [48] was found in all isolates.
ARGs varied between the different species, just like with iMGEs. For example, blaTEM-1B and dfrA17 were more prevalent in E. coli, while fosA, OqxA, and OqxB were more common in K. pneumoniae. There were only minor differences in the ARG profile between MLST ST. E. faecalis was the exception where st480 harboured macrolide resistance gene dfrG and lincosamide resistance gene lnu(B) compared to st6 and st774, which tended to carry the aminoglycoside resistance gene ant(6)-Ia, as can be seen in S5 Fig. Only minor differences in ARG prevalence between continents, as seen in S6 Fig.
Identification of transferable ARGs
The contigs were classified as being of either plasmid or chromosomal origin based on the consensus prediction of PlasClass and Platon. S. aureus had significantly lower estimated plasmid content than the other species (P-value 2.36 × 10−129; statistics in S3 Table), averaging 15.9 Kbp plasmid DNA per genome (~0.6% of the assembled DNA). E. coli, K. pneumoniae, and E. faecalis had comparable plasmid content relative to the total assembled DNA (~2.3%, ~ 3.8%, and ~2.7%), S7 Fig.
ARGs mobilised by plasmids were identified as they could spread to other bacteria. The number of mobilised ARGs varied between the species, with plasmid-borne ARGs being more common in Gram- bacteria than in Gram+ ones (Fig 3). Most ARGs in E. faecalis were located on the chromosome, except for Amphenicol and Glycopeptide resistance genes. S. aureus carried most Lincosamide, Streptogramin b, and tetracycline resistance genes on plasmids (S8 Fig). The Gram- species carried most ARGs on plasmids regardless of AMR class (Fig 3) except for the blaSHV family of beta-lactamases in K. pneumoniae that were primarily located on chromosomal contigs (S2 Appendix).
ARGs mobilised by iMGEs were also identified as they could facilitate the spread of resistance. Of the 7,159 predicted ARGs, 327 were carried by iMGEs of different types. S. aureus and E. coli had the highest abundance of iMGEs mobilised ARGs (Fig 4).
Hatched bars denote that the ARG is located on a plasmid.
Beta-lactamases were commonly associated with iMGEs in all species except E. faecalis. In S. aureus blaZ was mainly carried by the Tn552 unit-transposon, but in six cases it was found interspersed between copies of the IS ISSep1 or ISSau6. These MGEs could potentially form a comTn.
Various blaTEM beta-lactamases were carried by Tn2 and Tn801 unit-transposons in E. coli and K. pneumoniae. Other beta-lactamase genes, such as blaKPC-2, blaCTX-M-15, and blaSHV-1, were instead located between copies of ISEc9, ISEc26, and ISKpn31 that potentially could transpose the genes by forming a comTn. Despite E. faecalis harbouring a larger diversity and higher relative abundance of iMGEs (Fig 4) than the other species, only three iMGE were identified to mobilise ARGs. These ARGs and iMGEs were: erm(B) carried by Tn917, lnu(G) carried by Tn6260, and the Aminoglycoside resistance gene aac(6’)-aph(2’‘) harboured by an IS256-based comTn.
iMGEs associated ARGs indicate cross-species transposition
We found evidence of 21 MGEs and ARGs being transposed as a unit because they were repeatedly found at the same distance in multiple species in either the primary or secondary dataset. Eight of these pTUs were found on both plasmid and chromosomal contigs in two of the species, suggesting intra-cellular mobility. The secondary dataset was searched to confirm whether the pTUs were present in additional species beyond the four species in the primary dataset. Most pTUs were found in additional genera, further supporting the hypothesis that they have been transposed as a conserved unit. For example, ISEc9-blaCTX-M-15 was found in multiple genera, including Enterobacter, Escherichia, Klebsiella, and Serratia (Fig 5). This indicates the ability of MGEs to transmit genes between bacteria of clinical importance.
Rows are the pTUs; columns show the genera of the additional ~2000 isolates. The top bar display species stratified into Gram+ (white) and Gram- (black).
Species-specific differences in the MGE population
A total of 19,752 iMGEs were identified in all four species. Gram- bacteria, E. coli and K. pneumoniae, were found to harbour a larger abundance and diversity of known iMGEs (Fig 1 and S3 Fig) compared to the Gram+ S. aureus and E. faecalis. IS was the most common type of iMGEs in all species, but the proportion of plasmid-borne elements was higher in Gram- bacteria (Fig 6). MITEs and IME types of iMGEs were only found in E. coli and K. pneumoniae.
The number of predicted iMGE is normalized to the number of species isolates. The bar colours show the number of MGEs predicted to be plasmid or chromosome borne. Plasmid annotation is based on annotations from plasmid prediction software.
There was a systematic difference in the iMGEs population where most elements were only found in the same species, as can be seen in S9 Fig. For example, insertion sequences ISEc1, IS3, and IS609 were exclusively found in E. coli, while ISKpn38, ISKox1, and ISEc15 were only predicted in K. pneumoniae (S3 Appendix). Similarly, ISAu6 and ISSep3 and the transposon Tn554 were only found in S. aureus, while the insertion sequences ISLmo19, IS1062, and Tn6260 were only identified in E. faecalis (S3 Appendix).
iMGE profile corresponds with lineage
Isolates of the same ST tended to carry similar iMGEs. The relationship between iMGE profile and lineage was most pronounced in E. coli, where several STs (st3, st4, and st53) corresponded with isolated iMGE clusters. The relationship was also strong for S. aureus, K. pneumoniae, and E. faecalis. The clusters are presented in S10-S14 Figs.
Some lineages tended to share similar iMGE profiles. For instance, st506 and st43 in E. coli, st258 and st11 in K. pneumoniae, and st5 and st45 in S. aureus are samples with similar iMGE profiles. Some STs (st1 in E. coli and st11 in K. pneumoniae) consisted of multiple iMGE clusters, indicating possible strain-specific differences.
iMGEs shared by multiple species
A total of 102 iMGEs were found in multiple species, with four being common between S. aureus and E. faecalis and 98 by E. coli and K. pneumoniae. Of the four iMGEs shared by Gram+ species were two IS (IS256 and ISLmo19), one Tn (Tn5405), and one ICE (Tn6009), S14 Fig. IS was also the most frequent iMGE type found in E. coli and K. pneumoniae (84 elements), followed by Tns (6 elements). The Tns includes known ARG-carrying elements from the Tn2 (Tn2, Tn1000, Tn4656) and Tn3 (Tn5403) families. Only two ICEs (ICEKpnHS11286 and ICEEcoED1a) were identified in both species.
Differences in iMGE host ranges
The expanded isolate collection was analysed to investigate species-specific iMGE differences further. Most iMGEs were only found in related taxa, with ~39.2% of the iMGEs found in isolates of the same species, ~ 21.3% within the same genera, and ~19.9% within the same family (Fig 7). Interestingly, ~ 8.5% of the iMGEs were found in species of the same taxonomic class, and four iMGEs in multiple phyla. These included three IS (IS6100, ISSen9, and ISEfa8) and one Tn (Tn6082). The confinement of all iMGEs is presented in S4 Appendix.
The relationship between iMGEs and taxonomy was explored by creating a network that linked species together based on shared iMGEs. The network showed that related species were closely grouped in the graph as seen in S15 Fig. Four communities of densely connected species were identified with the Louvain method (Fig 8). These communities consisted of taxonomic families of the same class and Gram-staining group. For example, communities 1 and 2 consisted primarily of Gammaproteobacteria; community 1 contained Gram- and community 2 Gram+ families. However, Listeriaceae and Corynebacteriaceae were exceptions because they shared iMGEs with species of opposite gram-staining (ISLmo18 by Listeriaceae and IS6100, Tn6082 by Corynebacteriaceae).
Communities are clusters within the graph of species densely connected by iMGEs. The bars depict the number of species coloured by their taxonomic family. Hatched bars indicate Gram- species.
Discussion
The increasing prevalence of AMR poses a significant threat to global public health by contributing to increased mortality rates and the cost of healthcare [1,2]. MGEs play a critical role in the acquisition and dissemination of resistance genes, thereby facilitating the global spread of AMRs[6].
Here we present a comprehensive analysis of the global resistome and mobilome of four clinically significant pathogens using routine diagnostic samples collected without selection bias. This data offers an unbiased snapshot of infections occurring worldwide during 2020 [47].
Our findings revealed systematic differences in the abundance and diversity of ARGs among the four pathogens. The prevalence of certain beta-lactamase gene families and individual gene variants varied geographically, with some globally widespread while others were more prevalent in certain regions. Our findings are consistent with previous reports that the prevalence of extended-spectrum beta-lactamases (ESBL) in clinical isolates varies between different parts of the world [49]. We found the ESBL genes, such as blaOXA and blaCTX-M were more prevalent in isolates from Africa and Asia. This includes important ESBL genes such as blaCTX-M-15 and blaOXA-1, consistent with reports of high prevalence of ESBL in developing South Asian countries [50]. Conversely, blaSHV-106 harboured by K. pneumoniae, once thought to be geographically constrained to Portugal [51], was detected across multiple continents. Our findings are supported by reports of blaSHV-106 in Italy [52], Romania [53], and China [54], which underscores the ongoing dissemination of ESBL genes and emphasizes the need for global AMR surveillance efforts.
We investigated the mechanisms of ARG mobility and found several genes were mobilised by plasmids and iMGEs. The number of mobilised ARGs varied depending on the bacterial species, with E. faecalis having the fewest. Perhaps having intrinsic resistance to common antibiotics [48,55] has resulted in E. faecalis being less dependent on MGEs than species with less intrinsic resistance, such as E. coli and S. aureus.
We also found evidence of 21 ARGs potentially mobilised by nearby iMGEs, with eight of these units located on chromosomal and plasmid contigs across multiple species, indicating intracellular transposition events. Many pTUs contained iMGEs previously linked to ARG mobilisation. For example, IS6100 has been shown to mobilize resistance genes or promote genome rearrangements in Sphingomonas [56], Salmonella [57] and Citrobacter freundii [58]. Similarly, ISEc9 was frequently associated with blaTEM-1B and blaCTX-M-15 in pTUs found in Escherichia, Klebsiella, and Enterobacter, consistent with previous reports linking ISEc9 to members of the blaCTX-M family [59–61]. Notably, ISEc9 has been shown to facilitate the transformation of large DNA fragments with resistance genes between clinically relevant bacteria [62] indicating a potential to facilitate horizontal gene transfer independently of conjugating elements.
We also examined the broader distribution of iMGEs across species. E. coli exhibited the highest diversity of both iMGEs and ARGs, possibly reflecting its large accessory genome [63]. Smaller iMGEs like MITEs and IS elements were more abundant than larger transposons and conjugative elements (ICE, IME), a pattern consistent with our previous investigation of zoonotic Salmonella enterica [38]. Interestingly, we observed no apparent geographic influence on MGE diversity [38], contrasting with our findings from analysing the mobilome of sewage [64].
Instead, we found that taxonomically related species tended to share similar iMGEs profiles, suggesting the distribution to be influenced by bacterial lineage and host phylogeny. This lineage-specific pattern, consistent with a previous study of zoonotic Salmonella enterica [38], indicates strong vertical inheritance. However, the detection of certain iMGEs across multiple phyla, including both Gram- and Gram+ species, indicates that some elements have a broad host range. These findings align with our previous studies on ARG dissimilation pathways [64–66] and suggest that such iMGEs may be particularly important in spreading resistance genes across taxonomic boundaries. Incorporating surveillance of highly transmissible MGEs into existing antimicrobial resistance frameworks may improve early detection of ARGs with high risk of being transmitted and could guide prevention strategies.
To investigate potential transmission pathways, we applied Louvain’s community detection algorithm to model species connectivity based on shared iMGEs. This analysis revealed four clusters, primarily composed of bacteria of the same taxonomic class, and to a lesser extent, of the same gram-staining group. These clusters suggest gene transfer may be more likely within phylogenetically related species. While Louvain’s method has proven effective in analysing biological data [67] further research is needed to determine whether such interconnections correspond with increased risk of ARG transmission.
The study was limited by our data collection and sequencing methodology. The dataset was collected in 2020 during the COVID-19 pandemic, which may have influenced the diversity and prevalence of circulating pathogens. Although the samples represented 35 countries across multiple continents, broader geographic coverage or extending the sampling period would likely provide a more comprehensive representation of the global bacterial diversity.
Technical limitations also affected our ability to characterise the MGEs fully. The use of short-read sequencing limited our ability to resolve large and complex MGEs, such as ICEs and IMEs, likely leading to their underrepresentation. To mitigate classification bias, we used the two complementary plasmid prediction tools, PlasClass and Platon, which rely on different methodologies [39,40]. Their consensus prediction aligned well with known genomic context, for example, chromosomally encoded ARGs such as mecA in S. aureus [68] and Lsa efflux pumps in E. faecalis [48] were found on chromosomal contigs. Likewise, frequently mobilized ARG families like blaCTX-M, blaTEM, and blaOXA [69–71] were predominantly located on plasmids. In addition, we expanded the MobileElementFinder database with 1,686 IS and 70 Tns to improve iMGE prediction sensitivity. Despite these limitations, the study provides valuable insight into the global resistome and mobilome of clinically relevant bacteria.
Conclusion
MGEs can spread ARGs through multiple mechanisms, from direct carriage on plasmids to being integrated within a pTU. Our findings suggest that the transmissibility of iMGEs in pathogenic bacteria is primarily influenced by the bacterial lineage and host phylogeny. However, we identified 45 iMGEs present in species across multiple taxonomic classes, and four across multiple phyla, indicating that some elements may have a broad host range and a greater potential to disseminate ARGs across diverse bacterial populations. Additionally, we identified 21 genomic regions containing resistance genes potentially mobilised by MGEs, highlighting their role in gene transfer.
Mapping such mobilizable regions can help identify ARGs that could be transmitted. A deeper understanding of MGE transmissibility could enhance risk assessment and possibly inform future surveillance strategies. Future studies should aim to include a broader diversity of species to represent the clinical pathogens and the bacterial background better. Leveraging publicly available genomes and incorporating long-read sequencing are essential for resolving large and complex MGEs (e.g. plasmids, IMEs, and ICEs) and advancing our understanding of how these elements interact and evolve.
Supporting information
S1 Appendix. Sample information including accession numbers.
https://doi.org/10.1371/journal.pone.0330304.s001
(XLSX)
S2 Appendix. ARG composition for E. coli and K. pneumoniae and their predicted genomic location.
Genes are coloured by the class of antibiotics they yield resistance to.
https://doi.org/10.1371/journal.pone.0330304.s002
(XLSX)
S3 Appendix. Name, group and familiy information of identified integrated MGEs.
https://doi.org/10.1371/journal.pone.0330304.s003
(XLSX)
S4 Appendix. List of predicted iMGEs and in which species they were identified.
https://doi.org/10.1371/journal.pone.0330304.s004
(XLSX)
S1 Fig. Frequency of MLST sequence types per species in the dataset.
Rare sequence types for a given species (frequency < 5%) were amalgamated into the other category. The rare sequence type constituted ~32%−80%. Map was created using shapefiles from Natural Earth.
https://doi.org/10.1371/journal.pone.0330304.s005
(TIF)
S2 Fig. The number samples and location of participating diagnostic units.
a) Locations where the bacterial isolates were collected. The marker shape and colour designate the geographical region.; b) barplot showing the number of isolates per bacterial species and region.
https://doi.org/10.1371/journal.pone.0330304.s006
(TIF)
S3 Fig. Shannon diversity of iMGEs and ARGs per species.
https://doi.org/10.1371/journal.pone.0330304.s007
(TIF)
S4 Fig. Barplot displaying the relative abundance of ARG genes aggregated by antibiotic class per species.
https://doi.org/10.1371/journal.pone.0330304.s008
(TIF)
S5 Fig. Principal component of the ARG abundance for each species.
The scores are colored by MLST ST. The ten most common STs are displayed, and the rest are amalgamated into the other category. Ellipses shows the 90% confidence of score position.
https://doi.org/10.1371/journal.pone.0330304.s009
(TIF)
S6 Fig. Principal component of the ARG abundance for each species with scores colored by geographical region.
Ellipses show the 90% confidence of score position.
https://doi.org/10.1371/journal.pone.0330304.s010
(TIF)
S7 Fig. Boxplot showing the distribution of plasmid content for each species.
https://doi.org/10.1371/journal.pone.0330304.s011
(TIF)
S8 Fig. Predicted ARGs and their predicted genomic location for the Gram+ species E. faecalis and S. aureus.
Genes are coloured by the class of antibiotics they yield resistance to.
https://doi.org/10.1371/journal.pone.0330304.s012
(TIF)
S9 Fig. Principal component of the iMGE abundance coloured by species.
https://doi.org/10.1371/journal.pone.0330304.s013
(TIF)
S10 Fig. Dendrogram showing similarity of iMGE profile of E. coli isolates.
The inner circle shows MLST sequence types (ST) for the ten most frequent STs. Rare STs are amalgamated into the other category, and novel STs were omitted, showing gaps. The outer circle denotes the geographical region from which the isolate was collected. Isolates were clustered using Jaccard distance and average linkage.
https://doi.org/10.1371/journal.pone.0330304.s014
(TIF)
S11 Fig. Dendrogram showing similarity of iMGE profile of K. pneumoniae isolates.
The inner circle shows MLST sequence types (ST) for the ten most frequent STs. Rare STs are amalgamated into the other category, and novel STs were omitted, showing gaps. The outer circle denotes the geographical region from which the isolate was collected. Isolates were clustered using Jaccard distance and average linkage.
https://doi.org/10.1371/journal.pone.0330304.s015
(TIF)
S12 Fig. Dendrogram showing similarity of iMGE profile of S. aureus isolates.
The inner circle shows MLST sequence types (ST) for the ten most frequent STs. Rare STs are amalgamated into the other category, and novel STs were omitted, showing gaps. The outer circle denotes the geographical region from which the isolate was collected. Isolates were clustered using Jaccard distance and average linkage.
https://doi.org/10.1371/journal.pone.0330304.s016
(TIF)
S13 Fig. Dendrogram showing similarity of iMGE profile of E. faecalis isolates.
The inner circle shows MLST sequence types (ST) for the ten most frequent STs. Rare STs are amalgamated into the other category, and novel STs were omitted, showing gaps. The outer circle denotes the geographical region from which the isolate was collected. Isolates were clustered using Jaccard distance and average linkage.
https://doi.org/10.1371/journal.pone.0330304.s017
(TIF)
S14 Fig. The number of iMGEs found in either Gram+ or Gram- species.
https://doi.org/10.1371/journal.pone.0330304.s018
(TIF)
S15 Fig. Network graph showing bacterial species being connected by shared iMGEs.
Nodes represent species and are colored by family. Edges connecting nodes represent a shared iMGE.
https://doi.org/10.1371/journal.pone.0330304.s019
(TIF)
S1 Table. Number of beta-lactamase genes per continent in E. coli.
https://doi.org/10.1371/journal.pone.0330304.s020
(DOCX)
S2 Table. Number of beta-lactamase genes per continent in K. pneumoniae.
https://doi.org/10.1371/journal.pone.0330304.s021
(DOCX)
S3 Table. Statistics for two-way ANOVA testing difference in predicted plasmid content per species.
https://doi.org/10.1371/journal.pone.0330304.s022
(DOCX)
References
- 1.
World Health Organization. Antimicrobial resistance: Global Report on Surveillance. 2014 [cited 4 May 2024]. Available from: https://apps.who.int/iris/bitstream/10665/112642/1/9789241564748_eng.pdf
- 2. Robinson TP, Bu DP, Carrique-Mas J, Fèvre EM, Gilbert M, Grace D, et al. Antibiotic resistance is the quintessential One Health issue. Trans R Soc Trop Med Hyg. 2016;110(7):377–80. pmid:27475987
- 3. Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet. 2022;399(10325):629–55. pmid:35065702
- 4. Rice LB. Federal funding for the study of antimicrobial resistance in nosocomial pathogens: no ESKAPE. J Infect Dis. 2008;197(8):1079–81. pmid:18419525
- 5.
World Health Organization. Global Priority List of Antibiotic-Resistance Bacteria to Guide Research, Discovery, and Development of New Antibiotics. 2017 [cited 9 Jun 2023]. Available from: https://remed.org/wp-content/uploads/2017/03/lobal-priority-list-of-antibiotic-resistant-bacteria-2017.pdf
- 6. Partridge SR, Kwong SM, Firth N, Jensen SO. Mobile Genetic Elements Associated with Antimicrobial Resistance. Clin Microbiol Rev. 2018;31(4):e00088-17. pmid:30068738
- 7. Dionisio F, Zilhão R, Gama JA. Interactions between plasmids and other mobile genetic elements affect their transmission and persistence. Plasmid. 2019;102:29–36. pmid:30771401
- 8. Guédon G, Libante V, Coluzzi C, Payot S, Leblond-Bourget N. The Obscure World of Integrative and Mobilizable Elements, Highly Widespread Elements that Pirate Bacterial Conjugative Systems. Genes (Basel). 2017;8(11):337. pmid:29165361
- 9. Botelho J, Schulenburg H. The Role of Integrative and Conjugative Elements in Antibiotic Resistance Evolution. Trends Microbiol. 2021;29(1):8–18. pmid:32536522
- 10. Johnson CM, Grossman AD. Integrative and Conjugative Elements (ICEs): What They Do and How They Work. Annu Rev Genet. 2015;49:577–601. pmid:26473380
- 11. Siguier P, Gourbeyre E, Chandler M. Bacterial insertion sequences: their genomic impact and diversity. FEMS Microbiol Rev. 2014;38(5):865–91. pmid:24499397
- 12. Nzabarushimana E, Tang H. Insertion sequence elements-mediated structural variations in bacterial genomes. Mob DNA. 2018;9:29. pmid:30181787
- 13. Harmer CJ, Pong CH, Hall RM. Structures bounded by directly-oriented members of the IS26 family are pseudo-compound transposons. Plasmid. 2020;111:102530. pmid:32871211
- 14. Haniford DB, Ellis MJ. Transposons Tn10 and Tn5. Microbiol Spectr. 2015;3(1):631–45. pmid:26104553
- 15. Hansson K, Sundström L, Pelletier A, Roy PH. IntI2 integron integrase in Tn7. J Bacteriol. 2002;184(6):1712–21. pmid:11872723
- 16. Santos C, Caetano T, Ferreira S, Mendo S. Tn5090-like class 1 integron carrying bla(VIM-2) in a Pseudomonas putida strain from Portugal. Clin Microbiol Infect. 2010;16(10):1558–61. pmid:20121819
- 17. Fonseca ÉL, Vicente AC. Integron Functionality and Genome Innovation: An Update on the Subtle and Smart Strategy of Integrase and Gene Cassette Expression Regulation. Microorganisms. 2022;10(2):224. pmid:35208680
- 18. Lu M, Gong T, Zhang A, Tang B, Chen J, Zhang Z, et al. Mobile Genetic Elements in Streptococci. Curr Issues Mol Biol. 2019;32:123–66. pmid:31166171
- 19. Noel HR, Petrey JR, Palmer LD. Mobile genetic elements in Acinetobacter antibiotic‐resistance acquisition and dissemination. Annals of the New York Academy of Sciences. 2022;1518(1):166–82.
- 20. Malachowa N, DeLeo FR. Mobile genetic elements of Staphylococcus aureus. Cell Mol Life Sci. 2010;67(18):3057–71.
- 21. Horne T, Orr VT, Hall JP. How do interactions between mobile genetic elements affect horizontal gene transfer?. Curr Opin Microbiol. 2023;73:102282. pmid:36863168
- 22. Frost LS, Leplae R, Summers AO, Toussaint A. Mobile genetic elements: the agents of open source evolution. Nat Rev Microbiol. 2005;3(9):722–32. pmid:16138100
- 23. Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods. 2020;17(3):261–72.
- 24. Wein T, Hülter NF, Mizrahi I, Dagan T. Emergence of plasmid stability under non-selective conditions maintains antibiotic resistance. Nat Commun. 2019;10(1):2595. pmid:31197163
- 25.
Sidsel N, Gunhild L, Judit S, Laura B, Gábor G, Wojchiech C, et al. Whole genomes from bacteria collected at diagnostic units around the world 2020. 2022. https://doi.org/10.11583/DTU.21758456
- 26.
Center for Genomic Epidemiology. FoodQCPipeline. [cited 4 May 2024]. Available from: https://bitbucket.org/genomicepidemiology/foodqcpipeline/src/master/.
- 27.
Andrews S. FastQC A Quality Control tool for High Throughput Sequence Data. Babraham Bioinformatics; 2010 [cited 3 Jun 2015]. Available from: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/.
- 28. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol. 2012;19(5):455–77. pmid:22506599
- 29. Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013;29(8):1072–5. pmid:23422339
- 30. Hasman H, Saputra D, Sicheritz-Ponten T, Lund O, Svendsen CA, Frimodt-Møller N, et al. Rapid whole-genome sequencing for detection and characterization of microorganisms directly from clinical samples. J Clin Microbiol. 2014;52(1):139–46. pmid:24172157
- 31. Silva M, Machado MP, Silva DN, Rossi M, Moran-Gilad J, Santos S, et al. chewBBACA: A complete suite for gene-by-gene schema creation and strain identification. Microb Genom. 2018;4(3):e000166. pmid:29543149
- 32. Leopold SR, Goering RV, Witten A, Harmsen D, Mellmann A. Bacterial whole-genome sequencing revisited: portable, scalable, and standardized analysis for typing and detection of virulence and antibiotic resistance genes. J Clin Microbiol. 2014;52(7):2365–70. pmid:24759713
- 33. Zhou Z, Alikhan N-F, Mohamed K, Fan Y, Agama Study Group, Achtman M. The EnteroBase user’s guide, with case studies on Salmonella transmissions, Yersinia pestis phylogeny, and Escherichia core genomic diversity. Genome Res. 2020;30(1):138–52. pmid:31809257
- 34. Toorop MMA, Kraakman MEM, Hoogendijk IV, van Prehn J, Claas ECJ, Wessels E, et al. A core-genome multilocus sequence typing scheme for the detection of genetically related Streptococcus pyogenes clusters. J Clin Microbiol. 2023;61(11):e0055823. pmid:37815371
- 35.
Seemann T. mlst. [cited 24 May 2024]. Available from: https://github.com/tseemann/mlst
- 36. Jolley KA, Maiden MCJ. BIGSdb: Scalable analysis of bacterial genome variation at the population level. BMC Bioinformatics. 2010;11:595. pmid:21143983
- 37. Bortolaia V, Kaas RS, Ruppe E, Roberts MC, Schwarz S, Cattoir V, et al. ResFinder 4.0 for predictions of phenotypes from genotypes. J Antimicrob Chemother. 2020;75(12):3491–500. pmid:32780112
- 38. Johansson MHK, Bortolaia V, Tansirichaiya S, Aarestrup FM, Roberts AP, Petersen TN. Detection of mobile genetic elements associated with antibiotic resistance in Salmonella enterica using a newly developed web tool: MobileElementFinder. J Antimicrob Chemother. 2021;76(1):101–9. pmid:33009809
- 39. Pellow D, Mizrahi I, Shamir R. PlasClass improves plasmid sequence classification. PLoS Comput Biol. 2020;16(4):e1007781. pmid:32243433
- 40. Schwengers O, Barth P, Falgenhauer L, Hain T, Chakraborty T, Goesmann A. Platon: identification and characterization of bacterial plasmid contigs in short-read draft assemblies exploiting protein sequence-based replicon distribution scores. Microb Genom. 2020;6(10):mgen000398. pmid:32579097
- 41. Hunter JD. Matplotlib: A 2D Graphics Environment. Comput Sci Eng. 2007;9(3):90–5.
- 42. Waskom M. seaborn: statistical data visualization. J Open Source Softw. 2021;6(60):3021.
- 43. Titus Brown C, Irber L. sourmash: a library for MinHash sketching of DNA. JOSS. 2016;1(5):27.
- 44. Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, et al. BLAST+: architecture and applications. BMC Bioinformatics. 2009;10:421. pmid:20003500
- 45.
Hagberg AA, Schult DA, Swart PJ. Exploring Network Structure, Dynamics, and Function using NetworkX. In: Varoquaux G, Vaught T, Millman J, editors. Proceedings of the 7th Python in Science Conference. Pasadena; 2008. p. 11–15.
- 46. Blondel VD, Guillaume J-L, Lambiotte R, Lefebvre E. Fast unfolding of communities in large networks. J Stat Mech. 2008;2008(10):P10008.
- 47.
Nag S, Larsen G, Szarvas J, Gulyàs GM, Ciok W, Lagermann TMR, et al. Global collection of clinically relevant bacterial genomes from project TWIW 2020. 2023.
- 48. Hollenbeck BL, Rice LB. Intrinsic and acquired resistance mechanisms in enterococcus. Virulence. 2012;3(5):421–33. pmid:23076243
- 49. Pishtiwan AH, Khadija KM. Prevalence of blaTEM, blaSHV, and blaCTX-M Genes among ESBL-Producing Klebsiella pneumoniae and Escherichia coli Isolated from Thalassemia Patients in Erbil, Iraq. Mediterr J Hematol Infect Dis. 2019;11(1):e2019041. pmid:31308917
- 50. Husna A, Rahman MM, Badruzzaman ATM, Sikder MH, Islam MR, Rahman MT, et al. Extended-Spectrum β-Lactamases (ESBL): Challenges and Opportunities. Biomedicines. 2023;11(11):2937. pmid:38001938
- 51. Liakopoulos A, Mevius D, Ceccarelli D. A Review of SHV Extended-Spectrum β-Lactamases: Neglected Yet Ubiquitous. Front Microbiol. 2016;7:1374. pmid:27656166
- 52. Bonardi S, Cabassi CS, Fiaccadori E, Cavirani S, Parisi A, Bacci C, et al. Detection of carbapenemase- and ESBL-producing Klebsiella pneumoniae from bovine bulk milk and comparison with clinical human isolates in Italy. Int J Food Microbiol. 2023;387:110049. pmid:36521239
- 53. Surleac M, Czobor Barbu I, Paraschiv S, Popa LI, Gheorghe I, Marutescu L, et al. Whole genome sequencing snapshot of multi-drug resistant Klebsiella pneumoniae strains from hospitals and receiving wastewater treatment plants in Southern Romania. PLoS One. 2020;15(1):e0228079. pmid:31999747
- 54. Zhao H, He Z, Li Y, Sun B. Epidemiology of carbapenem-resistant Klebsiella pneumoniae ST15 of producing KPC-2, SHV-106 and CTX-M-15 in Anhui, China. BMC Microbiol. 2022;22(1):262. pmid:36319965
- 55. García-Solache M, Rice LB. The Enterococcus: a Model of Adaptability to Its Environment. Clin Microbiol Rev. 2019;32(2):e00058-18. pmid:30700430
- 56. Nagata Y, Endo R, Ito M, Ohtsubo Y, Tsuda M. Aerobic degradation of lindane (gamma-hexachlorocyclohexane) in bacteria and its biochemical and molecular basis. Appl Microbiol Biotechnol. 2007;76(4):741–52. pmid:17634937
- 57. Targant H, Doublet B, Aarestrup FM, Cloeckaert A, Madec J-Y. IS6100-mediated genetic rearrangement within the complex class 1 integron In104 of the Salmonella genomic island 1. J Antimicrob Chemother. 2010;65(7):1543–5. pmid:20483981
- 58. Liu N, Tang B, Wang H, Chen X, Wen P, Wang Z, et al. Coexistence of a novel NDM-1-encoding MDR plasmid and an IMP-4-encoding IncN-IncU hybrid plasmid in a clinical isolate of Citrobacter freundii BC73. Front Microbiol. 2024;15:1388651. pmid:39077736
- 59. Rossolini GM, D’Andrea MM, Mugnaioli C. The spread of CTX-M-type extended-spectrum beta-lactamases. Clin Microbiol Infect. 2008;14:33–41. pmid:18154526
- 60. Botts RT, Page DM, Bravo JA, Brown ML, Castilleja CC, Guzman VL, et al. Polluted wetlands contain multidrug-resistance plasmids encoding CTX-M-type extended-spectrum β-lactamases. Plasmid. 2023;126:102682.
- 61. Nakayama T, Yamaguchi T, Jinnai M, Kumeda Y, Hase A. ESBL-producing Vibrio vulnificus and V. alginolyticus harbour a plasmid encoding ISEc9 upstream of blaCTX-M-55 and qnrS2 isolated from imported seafood. Arch Microbiol. 2023;205(6):241. pmid:37198473
- 62. Domingues S, Lima T, Escobar C, Plantade J, Charpentier X, da Silva GJ. Large DNA fragment ISEc9-mediated transposition during natural transformation allows interspecies dissemination of antimicrobial resistance genes. Eur J Clin Microbiol Infect Dis. 2025;44(6):1417–24. pmid:40304893
- 63. Dobrindt U, Chowdary MG, Krumbholz G, Hacker J. Genome dynamics and its impact on evolution of Escherichia coli. Med Microbiol Immunol. 2010;199(3):145–54. pmid:20445988
- 64. Johansson MH, Aarestrup FM, Petersen T. Importance of mobile genetic elements for dissemination of antimicrobial resistance in metagenomic sewage samples across the world. PLoS One. 2023.
- 65. Hu Y, Yang X, Li J, Lv N, Liu F, Wu J, et al. The Bacterial Mobile Resistome Transfer Network Connecting the Animal and Human Microbiomes. Appl Environ Microbiol. 2016;82(22):6672–81. pmid:27613679
- 66. Ellabaan MMH, Munck C, Porse A, Imamovic L, Sommer MOA. Forecasting the dissemination of antibiotic resistance genes across bacterial genomes. Nat Commun. 2021;12(1):2435. pmid:33893312
- 67. Rahiminejad S, Maurya MR, Subramaniam S. Topological and functional comparison of community detection algorithms in biological networks. BMC Bioinformatics. 2019;20(1):212. pmid:31029085
- 68. Mlynarczyk-Bonikowska B, Kowalewski C, Krolak-Ulinska A, Marusza W. Molecular Mechanisms of Drug Resistance in Staphylococcus aureus. Int J Mol Sci. 2022;23(15):8088. pmid:35897667
- 69. Bradford PA. Extended-spectrum beta-lactamases in the 21st century: characterization, epidemiology, and detection of this important resistance threat. Clin Microbiol Rev. 2001;14(4):933–51, table of contents. pmid:11585791
- 70. D’Andrea MM, Arena F, Pallecchi L, Rossolini GM. CTX-M-type β-lactamases: a successful story of antibiotic resistance. Int J Med Microbiol. 2013;303(6–7):305–17. pmid:23490927
- 71. Evans BA, Amyes SGB. OXA β-lactamases. Clin Microbiol Rev. 2014;27(2):241–63. pmid:24696435