Abstract
Plasmids are major vehicles for the spread of antibiotic resistance genes. Some plasmids additionally carry virulence genes that enhance host pathogenicity. The convergence of resistance and virulence genes on the same plasmid poses significant risk, providing a mechanism to create pathogens that cause severe disease with limited treatment options. Colicin V (ColV)-like plasmids (ColVLPs) are virulence plasmids frequently carried by extra-intestinal pathogenic E. coli (ExPEC) that cause human and avian infection. Here, by generating and analysing a ColVLP database, we demonstrate that ColVLPs form four distinct sub-groups, characterised by genes encoding for Colicins V and M, with differing virulence and antimicrobial resistance gene carriage. Three of these sub-groups possess moderate-high resistance towards multiple antibiotic classes. We further describe ColVLP co-integrates that have acquired extensive resistance profiles, including against last line colistin, through recombination with co-resident plasmids. Using pMS7163A, a ColVLP from a virulent ExPEC strain, we also demonstrate that the ColVLP-encoded outer membrane protease virulence factor OmpTp works co-operatively with its chromosomal homolog to enhance ExPEC resistance against human cathelicidin (LL-37), an antimicrobial peptide expressed in the urinary tract. Together, our work characterises ColVLPs as high-risk mobile genetic elements that amplify the convergence of resistance and virulence in ExPEC.
Similar content being viewed by others
Introduction
Antibiotic resistance is a critical threat to global human health. Plasmids are a major driver contributing to this problem as they facilitate the horizontal dissemination of resistance genes via conjugation1. However, some plasmids are known to carry cargo genes encoding potent virulence factors2. The combined carriage of virulence and resistance determinants on the same plasmid has the potential to enhance the capacity of host bacteria to cause infections with reduced treatment options.
Colicin V (ColV) plasmids are a subset of Incompatibility (Inc) group F plasmids associated with enhanced virulence of extra-intestinal pathogenic Escherichia coli (ExPEC) that cause disease in animals and humans3,4. These plasmids are most frequently found in avian pathogenic E. coli (APEC) that cause infection in poultry5, but are also found in several ExPEC clones that cause human infections, including uropathogenic E. coli clones that cause urinary tract infection (UTI) and sepsis, such as ST1316, ST957,8 and ST589. ColV is a peptide antibiotic (microcin) that can kill or inhibit the growth of some E. coli and other closely related bacteria10. ColV production involves genes encoding its biosynthesis (cvaABC) and protection is conferred by an immunity protein (cvi) that prevents self-intoxification11,12. Although the cvaABC and cvi genes are markers of ColV plasmids, ColV production does not contribute to virulence13. Rather, increased virulence is associated with conserved cargo genes carried on ColV plasmids that include: sitABCD (metal transporter), shiF (aerobactin auxiliary factor), iucABCD/iutA (aerobactin siderophore), hlyF (upregulates outer membrane vesicle production), ompTp (involved in resistance to human cationic peptides), etsABC (ATP-binding cassette transporter), iss (involved in serum resistance) and iroBCDEN (salmochelin siderophore)4,14,15.
Recent studies have identified a group of plasmids closely related to ColV plasmids, termed ColB/M plasmids16. ColB/M plasmids carry most of the ColV-associated virulence cargo genes but not ColV itself6,7,16. The shared characteristics between ColB/M and ColV virulence plasmids have led to their redefinition collectively as ColV-like plasmids (ColVLPs), with the presence of ColV genes a marker (but not a strict requirement) for identification. Most ColVLP-encoded virulence loci are also found on the ExPEC chromosome within genomic islands, which include sitABCD, iucABCD/iutA, iss, iroBCDEN, ompT and shiF17,18,19. Prior studies have indicated sequence diversity between the ColVLP and chromosomal homologous genes4,20, but a comprehensive comparison of ColVLP and chromosomal homologues in terms of their evolutionary trajectory, virulence properties and potential synergistic effects is lacking. ExPEC that carry ColVLPs are associated with infections at the severe end of the UTI spectrum (e.g. acute pyelonephritis) and sepsis17. Most importantly, several studies have reported increased antimicrobial resistance (AMR) among E. coli strains carrying ColVLPs9,21, with some studies highlighting the mobile genetic element-mediated insertion of these AMR determinants on the ColVLP plasmid backbone—a concerning convergence of virulence and resistance on a conjugative mobile genetic element22.
Here, by collecting and analysing a database of completely sequenced ColVLPs, we describe a plasmid clustering profile associated with differences in AMR gene carriage as well as the existence of ColVLP co-integrates that carry resistance to last-line antibiotics. In terms of virulence, we demonstrate that the ColVLP-encoded outer membrane protease virulence factor OmpT (referred to in this study as OmpTp) contributes to ExPEC resistance to human cathelicidin (LL-37), a soluble antimicrobial peptide that protects against UTI23,24. Collectively, our work highlights the important roles of ColVLPs in augmenting ExPEC pathogenesis and resistance, an important advance given the high rates of infection caused by ExPEC clones that carry ColVLPs.
Results
ColVLPs can be split into six clusters based on gene content
We generated a dataset comprising 233 non-redundant, completely sequenced ColVLPs identified from PLSDB25 and NCBI RefSeq26 according to previously described criteria6. This stipulated that the ColVLPs contained >1 gene from at least four of the following sets: (i) cvaABC and cvi; (ii) iroBCDEN; (iii) iucABCD and iutA; (iv) etsABC; (v) ompT and hlyF; (vi) sitABCD. Plasmid taxonomic unit (PTU) classification based on average nucleotide identity revealed that most ColVLPs (225/233) belonged to, or were closely related to, PTU-FE, conjugative F plasmids that reside in E. coli27 (Supplementary Fig. 1).
Bacteria carrying these ColVLPs were isolated from multiple sources, with the most common based on available metadata being human (35%; 50/143), avian (29%; 42/143), porcine (18%; 26/143) and environmental (7%; 10/143) isolates (Fig. 1). Most ColVLPs were isolated from E. coli (96%; 224/233), and less frequently from S. enterica (6/233) and K. pneumoniae (3/233). Most ColVLPs contained more than one F replicon, with the most common combinations (using the FII:FIA:FIB formula28) being 18:-:1 (38%; 90/233), 24:-:1 (17%; 41/233), 2:-:1 (10%; 25/233) and 18:5,6:1 (5%; 12/233) (Fig. 1). When replicon carriage across the 233 ColVLPs was queried independently, the FIB_1 (85%; 199/233) and FII_18 (50%; 118/233) replicons were dominant. More than half of the ColVLPs (52%) carried all genes required for F plasmid conjugation and are potentially conjugative (Fig. 1), although transfer region polymorphisms and/or insertions may still impact transfer capability29.
Unique ORFs from 233 ColVLPs were combined to form a hypothetical plasmid. Using sequence similarity searches against the hypothetical plasmid at an 80% nucleotide sequence identity and alignment length threshold, a binary sequence indicating open-reading frame presence/absence for each plasmid was generated. Binary sequences were subjected to hierarchical clustering using Manhattan distance and visualised as a cladogram. Plasmids were arranged into clusters (n > 1) based on the total within sum of squares method, with minor (n ≤ 5) and major clusters (n > 5). Host isolate source, host species, replicon type, predicted conjugative ability and the presence of colicins, colicin immunity genes (imm), virulence and resistance features are annotated. Feature prevalence is illustrated in a bar plot above the metadata. Potential conjugative ability was predicted based on the presence of all genes required for F plasmid conjugation82 determined using BLASTn at a 90% nucleotide sequence identity and alignment length threshold. Replicons were identified using ABRicate v0.8 against the PlasmidFinder database using a 100% query length threshold. Virulence genes were identified using a BLASTn search at a 90% nucleotide sequence identity and 95% alignment length threshold, with operons being considered present only if all genes were individually identified. Antimicrobial resistance genes were identified using ABRicate v0.8 against the ARG-ANNOT database84 using a 100% query length threshold and grouped based on antibiotic class, with the following abbreviations: AGly aminoglycosides, Bla beta-lactamases, Fcyn fosfomycin, Flq fluoroquinolones, MLS macrolide-lincosamide-streptogramin, Phe phenicols, Rif rifampin, Sul sulphonamides, Tet tetracyclines and Tmt trimethoprim. Source data are provided as a Source data file in Supplementary Data 4.
Clustering of ColVLPs using an ORF-based analysis tool30,31 revealed two minor clusters (clusters 1 and 2) and four major clusters (clusters 3–6) (Fig. 1). A defining feature of the major clusters was the presence of colicins. Most ColVLPs of clusters 3 and 6 (>70%) encoded a complete ColV operon (encoding both the peptide antibiotic and the immunity protein), while ColVLPs of clusters 4 and 5 encoded a complete ColM operon (>80%) and the ColV and ColB immunity genes (>80%). The major clusters also differed in their virulence gene carriage, with >84% prevalence of all virulence loci (iroBCDEN, etsABC, ompTp, hlyF, iucABCD/iutA, shiF, sitABCD, iss and traT) in clusters 3 and 6, and lower virulence factor carriage for clusters 4 (iroBCDEN—76%; etsABC—17%; iucABCD/iutA—44%; shiF—43%; iss—75%; traT—81%) and 5 (etsABC—19%; iucABCD/iutA—69%; shiF—69%; traT—19%) (Fig. 1). The prevalence of AMR genes between these major clusters also differed. ColVLPs from cluster 3 had low carriage rates of all AMR genes examined, while ColVLPs of cluster 6 had moderate rates of resistance towards aminoglycosides, beta-lactams, sulphonamides, tetracyclines and trimethoprim antibiotics (~35–45%) (Fig. 1). ColVLPs from clusters 4 and 5 possessed high AMR gene carriage across multiple antibiotic classes including aminoglycoside (~65% and ~100%), beta-lactam (~75% and ~75%), fluoroquinolone (~25% and ~45%), macrolide (~25% and ~50%), phenicol (~35% and ~70%), tetracycline (~75% and ~55%) and trimethoprim (~55% and ~90%) antibiotics (Fig. 1). The presence/absence of each AMR gene determinant is detailed in Supplementary Data 4 and visualised in Supplementary Fig. 2, as the independent acquisition of each determinant can inform ColVLP evolutionary history, particularly in the case of rare determinants such as sul329. Taken together, the analysis revealed that ColVLPs can be broadly divided into two major groups: (i) clusters 4 and 5, characterised by the carriage of ColM and the association with lower virulence and higher AMR gene carriage, or (ii) clusters 3 and 6, characterised by the carriage of ColV and the association with higher virulence and low-moderate AMR gene carriage.
ColVLP variation includes the formation of co-integrate plasmids
Outside of the major clusters, our analysis also identified minor clusters 1 and 2, which consisted of three plasmids each that carried additional IncHI2/IncHI2A and IncN replicons, respectively (Supplementary Fig. 3). In addition to ColVLP-associated virulence loci and numerous AMR genes, these hybrid plasmids also carry mcr-1 (Cluster 1: 2/3; Cluster 2: 3/3; Fig. 1), conferring resistance to the last-line antibiotic colistin. A detailed analysis of these plasmids revealed two distinct lineages of ColVLP hybrids that contain regions from IncHI2-like plasmids (Supplementary Fig. 4), likely acquired from co-integration events. Cluster 1 ColVLPs (isolated from China) had significant sequence homology to the reference ColVLP pMS7163A32 and the IncHI2 plasmid pSTM6-27533 across regions encoding for replication, virulence and transfer (Fig. 2A; Supplementary Fig. 4). Cluster 2 ColVLPs (isolated in the USA) similarly had significant sequence homology to the aforementioned reference plasmids. However, these plasmids lacked an IncHI2 replicon and instead carried an IncN replicon, despite no significant homology to the IncN reference plasmid R46 (AY046276.1)34 (Fig. 2A; Supplementary Fig. 4).
A Pairwise comparisons of IncHI2/ColVLP co-integrate plasmids. Reference sequences were a concatenation of the reference ColVLP pMS7136A (NZ_CP026854.1) and the reference IncHI2 plasmid pSTM6-275 (NZ_CP019647.1). Pairwise comparisons were performed using BLASTn and visualised using EasyFig81 with a minimum alignment length of 1000 bp. Features are coloured as follows: IncF tra transfer region—blue; ColVLP virulence loci—red; IncHI2 tra1 transfer region—orange; IncHI2 tra2 transfer region—green; IS26—pink. B Split decomposition network of ColVLP co-integrates. Twenty non-redundant ColVLP co-integrates were identified from NCBI using the criteria of Liu et al.6 and PlasmidFinder replicons42. A network visualising plasmid relatedness was drawn using binary sequences generated using the open-reading-frame-based binarised structure network analysis tool. Clusters were determined based on a Dice index of >0.71. C Number of AMR genes (n) in ColVLP co-integrates and ColVLPs. Antimicrobial resistance genes were identified using ABRicate v0.8 against the ARG-ANNOT database84 using a 100% query length threshold. Data is presented as a boxplot, where the box limits represent first and third quartiles, the internal line represents the median, and whiskers represent data within a ±1.5 interquartile range. Significance was determined using an unpaired, two-tailed t-test. Figure acronyms: AMR antimicrobial resistance.
IS26, an insertion element associated with the mobilisation of resistance genes35, has been previously implicated in co-integrate formation36 and is known to be present in both ColVLPs22,29,37 and IncHI2-like plasmids33,38. Investigation of our co-integrate plasmids for IS26 revealed that all co-integrates harboured multiple IS26 copies (between 6-15 copies) in regions between the ColVLP and co-resident plasmid (Fig. 2A), consistent with previous reports that demonstrate the role of IS26 in co-integrate formation35,36,39,40.
ColVLP co-integrates carry high AMR gene loads
The presence of co-integrated ColVLPs prompted a wider search for additional, more recent chimeric ColVLPs outside of our collection. By querying NCBI for ColVLPs using the criteria of Liu et al.6 and PlasmidFinder replicons41, we identified 20 non-redundant ColVLPs which contained additional IncHI1, IncHI2, IncI1, IncN, IncX, or IncR replicons (Supplementary Data 1). Pairwise comparisons confirmed regions of high similarity to their respective reference plasmids, except for two IncX plasmids (AP017611.1 and CP0916721.1) (Supplementary Figs. 5–10). Of these 20 plasmids, 12 had available whole genome assembly data, allowing us to perform MLST classification of their host chromosomes. These ColVLP co-integrates were isolated from diverse E. coli STs, including ST101, ST141, ST93, ST23, ST457, ST744, ST359, ST1196, ST88 and ST117 (Supplementary Data 1). Importantly, these STs all belong to the ExPEC/APEC pathotype associated with human, avian and/or other animal infections42,43,44,45,46,47,48.
Sequence analysis of these 20 co-integrates revealed five clusters of closely related plasmids (largely corresponding to their co-integrant Incompatibility group): IncHI2+, IncN+, IncX1+, IncHI1+ and IncN/IncR+ (Fig. 2B). The five plasmids in the IncHI2+ cluster (of which two were mcr-1 positive) were isolated from human, avian, porcine and fish sources, suggestive of zoonotic transmission and/or conjugation between human and animal isolates. All IncN+ ColVLPs were mcr-1 positive and from human isolates. As these hybrid plasmids were the result of co-integration events, we hypothesised that they would also acquire the AMR repertoire of their co-integrant plasmid. Indeed, a comparison of the number of AMR genes revealed significantly more AMR genes on ColVLP co-integrates compared to their non-hybrid counterparts (p < 0.0001; Fig. 2C). Interestingly, all co-integrates except for one IncI1+ plasmid contained IS26 insertion elements, mostly located in regions between the ColVLP and co-resident plasmid (Supplementary Figs. 5–10), further supporting the role of IS26 in co-integrate formation. Overall, our identification of these ExPEC- and APEC-associated ColVLP co-integrates identified an underlying group of high-risk plasmids with the potential to confer both increased antibiotic resistance and enhanced virulence to E. coli with strong zoonotic potential.
ColVLP/AMR plasmid co-integration is likely driven by ST-specific plasmid carriage patterns
ColVLP co-integrates can arise when ColVLPs recombine with co-resident plasmids. To better understand this phenomenon, we first assessed the frequency at which ColVLPs are carried with additional plasmids by utilising a draft genome database consisting of 9962 E. coli genomes from the top 100 STs in Enterobase (100ST database)49 and querying for the presence of ColVLPs and additional non-IncF replicons, assuming that distinct replicons were located on different plasmid backbones. We found ColVLPs in most STs, with a strong association with human and avian isolates (Supplementary Fig. 11). Across all ColVLP+ strains in the 100ST database (1469/9962; Fig. 3A), we also found that most strains either carried the ColVLP alone (~40%) or carried one additional non-IncF replicon (~40%), although up to five additional non-IncF replicons were found in some strains (Fig. 3A, G). At the ST level, the distribution of additional replicons was ST-specific, with >50% of strains in some STs carrying ColVLPs alone (ST73, ST95, ST117, ST162, ST3580), and other STs with >50% of strains carrying ColVLPs with one or more additional non-IncF replicon (ST23, ST57, ST88, ST428, ST602) (Fig. 3B). Notably, ST131 was absent from the top 20 ColVLP+ STs, as ColVLP carriage is specific to the ST131 Clade B sub-lineage (55.5%; Supplementary Fig. 12).
Presence of additional replicons in ColVLP+ draft genomes of the A 100ST database and B top 10 ColVLP+ STs within the 100ST database. The 100ST database consists of 9969 draft E. coli genomes from the top 100 STs in Enterobase. Replicon co-carriage network in the C 100ST, D ST95 and E ST131 (Clade B) databases. The ST95 and ST131 (Clade B) databases consist of 2118 and 344 draft genomes, respectively. ColVLPs were plotted as a separate node from IncF plasmids (IncF non-ColVLP). Total plasmid co-carriage events (n ≥ 10) were plotted using Cytoscape v3.9.1. Red lines denote the top four associations with ColVLPs. Line thickness between nodes is proportional to the number of co-carriage events. F Percent of ColVLP+ isolates in the 100ST, ST95 and ST131 Clade B databases. ColVLP+ isolates were tallied regardless of co-carriage events. Plasmid combinations of ColVLP+ isolates in the G 100ST, H ST95 and I ST131 (Clade B) databases. Plasmid combination frequency was normalised to the total number of ColVLP+ isolates in each database. The top six combinations of each database are shown. ColVLPs were identified using the criteria of Liu et al.6. Additional replicons were identified using ABRicate v0.8 against the PlasmidFinder42 database using a 100% query length threshold. Figure acronyms: ST sequence type. Source data are provided as a Source data file in Supplementary Data 5,7 and 8.
To investigate the replicons associated with ColVLP carriage, we plotted all significant co-carriage events in the 100ST database as a network. The top four replicons associated with ColVLP carriage were IncI1, IncX, IncB/O/K/Z and IncH (Fig. 3C). These co-carriage associations were also ST-dependent, as highlighted by the ColVLP-replicon associations of ST95 and ST131 (Clade B), two pandemic STs typically associated with ColVLP carriage, which were quantified with ST-specific databases of 2118 and 344 draft genomes, respectively. In ColVLP+ ST95 (1268/2118; Fig. 3F), the top four replicons associated with ColVLP carriage were (in order of magnitude): IncB/O/K/Z, IncX, IncH and IncI1 (Fig. 3D). In ColVLP+ ST131 Clade B genomes (191/344; Fig. 3F), the top four replicons were: IncI1, IncH, IncX and IncN (Fig. 3E). Notably, strains of ST131 Clade B had higher replicon diversity compared to ST95. ST131 Clade B strains, in addition to carrying ColVLPs alone (31%), frequently contained combinations of ColVLP + IncI1 (24%); + IncI1 and IncH (10%); or + IncX (7%) (Fig. 3I). In contrast, the majority of ST95 carried ColVLP alone (75%) or with an additional IncB/O/K/Z plasmid (13%) (Fig. 3H). Taken together, the data demonstrate that ColVLP+ E. coli strains frequently carry a diverse array of plasmid replicons, in part reflecting the existence of multiple co-resident plasmids in ST-specific patterns.
Virulence genes on ColVLPs are distinct from their chromosomal homologues
Several virulence determinants frequently associated with ColVLPs are also found on the chromosome, including the siderophores salmochelin (iroBCDEN) and aerobactin (iucABCD/iutA), the metal acquisition system (sitABCD), the increased serum survival gene (iss), the outer membrane protease (ompT) and the aerobactin auxiliary secretion factor (shiF)4,50,51. To investigate genetic differences between the chromosomal- and plasmid-encoded genes, we constructed phylogenetic trees of these virulence genes using E. coli chromosomal homologues identified from the NCBI RefSeq (1,377 chromosomes from completely sequenced strains) and their respective ColVLP-associated homologues from the 233 ColVLP database (Fig. 4A–F). For all virulence genes, most ColVLP-associated sequences clustered separately from their chromosomal homologues, indicating that the majority of ColVLPs carry a set of virulence loci distinct from that of their host chromosome.
Phylogeny of the virulence genes A iroBCDEN, B iucABCD/iutA, C shiF, D sitABCD, E iss and F ompT. Nucleotide sequences were extracted from the NCBI RefSeq (1,377 E. coli chromosomes) and 233 ColV-like plasmid (ColVLP) collection, aligned using Clustal Omega, and visualised as a midpoint-rooted tree using FastTree and the interactive Tree of Life. Potential recombinant homologues are highlighted in blue. G Nucleotide identity (%) comparing the non-conserved sites of the recombinant sitABCD operons with representative chromosomal and ColVLP homologues. Alignments between the four potential recombinant sitABCD operons (from NZ_CP019018.1, CP056868.1, NZ_CP032068.1, NZ_CP059908.1) and the representative homologues (ColVLP - NZ_CP026854.1, Chromosome - CP033884.1) were performed using CLC Main Workbench. Figure acronyms: SNP single nucleotide polymorphism, CHR chromosome.
We also noted rare instances where genes of ColVLP and chromosomal origin were interspersed in the phylogeny. To investigate these homologues for recombination, we manually checked the following for possible recombination: (i) every ColVLP-associated sequence which had a long branch indicative of large sequence variation; and (ii) every sequence that did not cluster with its associated group.
For the virulence genes iucABCD/iutA, shiF and ompT, all outliers were confirmed to be the result of inactivating mutations (frameshift or internal stop) or suspected mis-assembly. For iss, only one outlier was found, where a chromosomal-associated copy was found to cluster with ColVLP-associated homologues—while possibly recombination, more sequencing data is needed to confirm this observation.
For iroBCDEN and sitABCD, the ColVLP-associated homologues with long branches were also confirmed to contain inactivating mutations, and sitABCD additionally had two chromosomal homologues that clustered with ColVLP-associated copies due to suspected mis-assembly. Despite this, we were able to find convincing evidence of homologous recombination between ColVLPs and the chromosome. In the case of iroBCDEN, four fully intact operons that shared 99.97% nucleotide sequence similarity to a ColVLP-associated sequence were found on the E. coli chromosome (NZ_CP062985.1, NC_017632.1, NZ_CP012631.1, NZ_CP019777.1). More interestingly, with sitABCD, we found four ColVLP-associated sequences that formed their unique cluster within the chromosomal homologues. An alignment and comparison of non-conserved sites revealed that these four ColVLP-associated sitABCD sequences were hybrids between a ColVLP and chromosomal homologue, where the initial section (sitABC and start of sitD) shared high conservation with the reference chromosomal sequence, and the latter section of sitD shared high conservation with the reference ColVLP-associated sequence (Fig. 4G).
Taken together, our results indicate that while most virulence loci carried on the ColVLP are distinct from their chromosomal counterparts, homologous recombination events can occur, resulting in switching of the entire operon (as evidenced by iroBCDEN) or even a partial section of a virulence gene (as evidenced by sitD) between ColVLPs and the E. coli chromosome.
OmpT variants are functional but exhibit differences in substrate specificity
The phylogenetic distribution of the ColVLP-encoded and chromosomal-encoded virulence genes suggests conserved differences in their coding sequences. To explore if these differences altered function, we investigated OmpT, an outer membrane protease that cleaves antimicrobial peptides between two dibasic residues52,53 and promotes resistance to human LL-3754. In E. coli, two main OmpT variants exist—the chromosomal-encoded OmpT and the ColVLP-encoded OmpTp (also referred to as ArlC)19. A query of 1377 complete E. coli genomes from NCBI RefSeq26 revealed that ~65% of strains carry at least one ompT, with ~52.3% carrying a single copy on the chromosome and ~5% carrying at least one chromosomal and one plasmid copy (Fig. 5A). Notably, when the 100ST draft genomes was ranked for ompT carriage, >50% strains from the top five STs (ST428, ST117, ST23, ST88, ST141), and 39-50% of strains from the next five ranked STs (ST101, ST58, ST95, ST906, ST162), contained >1 ompT allele (Fig. 5B). These STs all belong to ExPEC/APEC pathotypes6,7,9,43,45,55,56,57,58,59.
A Predicted ompT genotype of 1377 E. coli assemblies. Prevalence of ompT in chromosomal and plasmid contigs was determined using the MS7163 ompT1 nucleotide sequence as a BLASTn query against 1377 complete E. coli genomes from NCBI RefSeq at a 65% query length and alignment threshold. B Top 10 STs carrying >1 ompT copy. Prevalence of ompT in the 100ST E. coli draft genomes database was determined using the MS7163 ompT1 nucleotide sequence as a BLASTn query. Superimposition of predicted C OmpT1 and OmpTp, D OmpT2 and OmpTp, E OmpT1 and OmpT2 protein structures from MS7163. Protein structures were predicted using ColabFold and annotated using ChimeraX. F FRET complement C2-derived peptide cleavage over time by MG1655 ompT::Cm expressing OmpT1, OmpT2, or OmpTp. All readouts were normalised against time 0 fluorescence. Data is presented as Mean ± SD of three biological replicates. One-way ANOVA with Dunnett’s multiple comparisons against the MG1655 ompT::Cm + pSU2718 control were performed on endpoint values. G Liquid chromatography–mass spectrometry (LC-MS) analyses of OmpT-mediated LL-37 cleavage. The LL-37 sequence is shown with predicted OmpT cleavage sites as a bold arrow. The bar graph shows the concentration of LL-3719-37 after incubation with MG1655 ompT::Cm + pOmpT1/2/p strains, quantified using LC-MS. Data are presented as Mean ± SD of three biological replicates. Significance was determined using a One-way ANOVA with Dunnett’s multiple comparisons against the pSU2718 vector-only control. H Survival of MS7163 derivatives against 0.05 mg/mL LL-37. Data are represented as Mean ± SD of five biological replicates. Significance was determined using a One-way ANOVA with Dunnett’s multiple comparisons against the WT strain. Survival of strains after 2 h of incubation with LL-37 is expressed as a percentage of the CFU count in PBS control at time 0. Figure acronyms: LPS lipopolysaccharide, WT wildtype, AA amino acid, AUC area under the curve. Source data are provided as a Source data file in Supplementary Data 5 and 6.
Using MS7163, an ST95 ExPEC reference strain that carries three ompT genes (two chromosomal-encoded alleles—ompT1 and ompT2, and one ColVLP-encoded allele—ompTp)32, we investigated the contribution of OmpTp to LL-37 resistance. The two chromosomal OmpT variants are highly similar (sharing 98.7% amino acid identity) and distinct from the plasmid variant (73.2% amino acid identity to either chromosomal copy) (Supplementary Fig. 13). A comparison of the active, catalytic and LPS-binding sites between OmpT1/OmpT2 and OmpTp revealed that all key residues were conserved, except for the two LPS-binding sites at residues 175 (R175H) and 226 (K226N) (Fig. 5C and Supplementary Fig. 13). Superimposition of the predicted OmpT1/2/p structures generated in ColabFold revealed almost identical structures, comprising a beta-barrel that extends from the outer membrane to the extracellular space (Fig. 5C–E).
To test for the activity of each OmpT variant, each individual allele was PCR amplified and cloned into the low-copy-number pSU2718-Km expression vector. This generated a series of constructs containing each individual ompT allele placed under the control of an identical promoter and ribosomal binding site. These resultant constructs were then used to transform an E. coli K-12 MG1655 mutant deleted for ompT (i.e. MG1655 ompT::Cm), with two assays subsequently performed to test the function of each ompT allelic variant. First, we used fluorescence resonance energy transfer (FRET) to measure the cleavage of a fluorescently labelled synthetic peptide derived from the sequence of complement C2, which revealed the greatest activity for OmpTp followed by OmpT1, but no detectable activity for OmpT2 (Fig. 5F). Second, we tested the ability of each OmpT variant to cleave LL-37 by utilising LC-MS to track a predicted LL-37 cleavage product. All OmpT variants were able to cleave LL-37 (Fig. 5G). Taken together, this data reveals that the ExPEC strain MS7163 encodes for three functional OmpT variants that exhibit variable substrate specificity, with OmpTp displaying activity against two different peptides.
The ColVLP OmpTp variant contributes to MS7163 LL-37 resistance
To investigate the contribution of the ColVLP-encoded OmpTp variant to LL-37 resistance, MS7163 ompT mutants expressing one, two or no active OmpT variants were generated using λ-Red-mediated homologous recombination. The survival of each mutant was examined after 2 h incubation with LL-37 in PBS. In these experiments, survival of the MS7163 triple mutant was significantly attenuated compared to the WT parent strain (p < 0.0001; Fig. 5H), demonstrating that the combined absence of all three OmpT variants enhances susceptibility to LL-37. The expression of any single OmpT variant did not restore LL-37 resistance to WT levels (OmpTp [Fig. 5H]; OmpT1 or OmpT2 [Supplementary Fig. 14A]), even when expressed from a plasmid expression vector (Supplementary Fig. 14B), revealing the requirement for a cooperative mode of resistance involving multiple OmpT variants. Despite this, the expression of OmpT1 and OmpTp consistently increased LL-37 resistance without always reaching statistical significance (Supplementary Fig. 14A; Fig. 5H). Notably, the co-expression of OmpTp+OmpT1 restored LL-37 resistance to a level comparable to WT (Fig. 5H), but this was not observed for the co-expression of OmpTp+OmpT2 nor OmpT1+OmpT2 (Supplementary Fig. 14A). Taken together, the data demonstrate a role for the ColVLP-encoded OmpTp variant in resistance to LL-37.
Discussion
Here we report on the convergence of virulence and antibiotic resistance in plasmids associated with ExPEC strains that cause human and animal infections. By generating a detailed ColVLP cladogram using an ORF-based binarised structure network analysis tool30, we identified clusters of closely related virulence plasmids and co-integrate plasmids that harbour high AMR gene burden. Furthermore, we show that the ColVLP-encoded OmpTp protease contributes to ExPEC resistance to LL-37, providing experimental evidence to link its function to enhanced virulence.
To date, most studies tracking ColVLP evolution have done so by tracing the acquisition and subsequent modification of mobile genetic elements carrying AMR determinants in a group of closely related FII-2/FIB-1-containing ColVLPs22,37. These molecular signatures include the presence of a class 1 integron truncated by IS2637, the acquisition of a Tn1721/Tn21 hybrid transposon, the acquisition and in situ modification of Tn1721, as well as the presence of Colicin Ia22. While these approaches allow for high-resolution understanding of plasmid evolution, they focus on sub-lineages of FII-2/FIB-1 ColVLPs (part of cluster 6; Supplementary Fig. 15). Here, by utilising an ORF-based approach independent of genetic markers and a conserved backbone, we broaden these existing analyses by providing a comprehensive, evolution-independent overview of currently sequenced ColVLPs (clustered by gene content). Our analysis identified diverse clusters of ColVLPs, providing a framework to characterise their contributions to ExPEC pathogenesis. We do note, however, that there are several limitations to our binary clustering methodology30. The methodology is annotation-dependent; independently acquired antibiotic resistance genes may influence clustering, and non-related plasmids may be clustered closer together than expected due to many shared 0’s between pairwise alignments that indicate no homology. Given these limitations, we suggest this method is not optimal for inferring plasmid evolution, but instead is better suited to clustering annotated and related plasmids, such as ColVLPs.
Our study revealed that ColVLPs are largely restricted to E. coli and are only rarely found in Klebsiella pneumoniae or Salmonella enterica. Given the conjugative nature of some ColVLPs, their restriction to E. coli suggests that ColVLPs have adapted to their E. coli hosts and may not be stably maintained when acquired by other Enterobacteriaceae. Despite this, there is a remarkable similarity between ColVLPs of ExPEC and virulence plasmids of hypervirulent Klebsiella pneumoniae (hvKp)60. Both plasmid types carry virulence genes encoding the salmochelin and aerobactin siderophore systems61 that contribute to the pathogenicity of their hosts17,62. In contrast, some virulence genes remain unique to each plasmid, such as rmpA (located on the hvKp virulence plasmid), whose expression enhances capsule production and results in a hypermucoid phenotype, and ompT (located on ColVLPs). In line with our findings, several completely sequenced hvKp virulence plasmids contain AMR genes likely acquired through recombination (mediated by IS elements) or co-integration63,64,65. It remains to be seen if a similar convergence of AMR and virulence determinants exists in hvKp virulence plasmids. Such convergence is alarming not only because of the increased impact these plasmids can have on both pathogenicity and resistance, but also because antibiotic use could select for hypervirulent strains harbouring these plasmids.
The identification of ColVLP co-integrates reveals a mechanism to explain how ColVLPs can rapidly expand their AMR repertoire and host range, a phenomenon observed in other studies66,67,68. While some studies demonstrate co-integration as a transient event between co-resident plasmids69, stable clinically-relevant plasmid co-integrates have been described, such as the enterotoxigenic E. coli virulence plasmid pCoo—a co-integrate between an IncF and an IncI1 plasmid70. The independent identification, sequencing and assembly of closely related mcr-1-positive co-integrates (Fig. 2) strongly suggests stability or repeated co-integration rather than transience, although this remains to be experimentally validated. Regardless, the identification of these mcr-1-positive ColVLP co-integrates from human ExPEC strains warrants deeper investigation through the implementation of unbiased (long-read) sequencing to capture the full extent of their distribution and transmission. Indeed, the problem of incomplete plasmid assemblies is a limitation of our plasmid replicon co-carriage analyses. Although our data show that ColVLP+ strains can carry additional replicons (IncI1, IncX, IncB/O/K/Z and IncH as the most common), we cannot differentiate between independent co-occurrence or co-occurrence on the same plasmid backbone without closed assemblies.
An intriguing aspect of our analyses was the observation that some ColVLP-encoded virulence loci have recombined with their host chromosome, resulting in full operon switching (iroBCDEN; Fig. 4A) to hybrid sequences (sitABCD; Fig. 4G). Here, our database consisted of completely sequenced E. coli genomes irrespective of ColVLP carriage and is likely predominantly from strains that do not normally carry ColVLPs. As recombination between the ColVLP and the chromosome can only occur when a strain carries a ColVLP, we suspect these recombination events occur at a rate higher than what our results suggest. To fully elucidate the extent of recombination, a large collection of completely sequenced genomes belonging to ColVLP+ STs is required, but is currently lacking.
Prior studies have suggested that the presence of multiple OmpT variants in the same strain confers a fitness advantage by expanding substrate range and protease activity against antimicrobial peptides, including LL-37 and RNase 7, that contribute to innate defence in the urinary tract19,71. Our results support the existence of distinct substrate preferences for each OmpT variant, noting also that each variant could effectively cleave LL-37 following plasmid-mediated expression in the K-12 host strain MG1655 (Fig. 5G). An important advance in our study was the demonstration that the expression of OmpTp and OmpT1 at wild-type levels enhanced protection of the ExPEC strain MS7163 against LL-37 (Fig. 5H), providing direct evidence that the ColVLP-encoded OmpTp plays a role in virulence. An underexplored aspect of OmpT-mediated LL-37 cleavage is the putative requirement of LPS for OmpT activity72. Although a recent study demonstrated that mutation of the predicted LPS-binding sites did not abolish OmpT activity73, this does not rule out the possibility for LPS non-specific binding or additional LPS-binding sites. While we identified two residues involved in OmpT-LPS binding that differed between OmpT1/OmpT2 and OmpTp (R175H and K226N), the impact of these amino acid changes on substrate specificity remains to be determined. Overall, the findings from our study, together with a previous report that showed OmpTp can cleave the antimicrobial peptide RNase 719, demonstrate that the ColVLP-encoded OmpTp protease functions in tandem with its OmpT chromosomal counterparts to enhance ExPEC survival in the presence of soluble innate immune effectors.
In summary, we have detailed the genetic characteristics of ColVLPs. These plasmids function as a horizontally disseminated vehicle to enhance the virulence armoury of ExPEC through the production of factors that lead to increased iron acquisition via siderophores, metal ion transport, serum resistance and the cleavage of host antimicrobial peptides. By demonstrating the carriage of multiple AMR genes on these plasmids, including co-integrate ColVLPs that harbour multiple replicons, our findings expand the role of ColVLPs in the plasmid-driven convergence of virulence and resistance in ExPEC.
Methods
Bacterial strains and growth conditions
All strains were routinely cultured at 37 °C in either lysogeny broth (LB) or N-minimal medium (50 mM Bis-Tris, 5 mM KCl, 7.5 mM (NH4)2SO4, 0.5 mM K2SO4, 0.5 mM KH2PO4, 0.1% casamino acids) supplemented with 0.2% glucose and 1 mM MgCl2 (NMM) under shaking (250 rpm) or static conditions. Antibiotics were supplemented at the following concentrations as appropriate: chloramphenicol (30 µg/mL); kanamycin (100 µg/mL), gentamycin (20 µg/mL). All strains were stocked at −80 °C in 15% glycerol. A full list of strains and plasmids used is available in Supplementary Data 2. MS7163 is a reference ST95 ColVLP+ ExPEC strain which was isolated from a patient with acute pyelonephritis and belongs to the virulent O45:K1:H7 clone that causes one-third of all neonatal meningitis cases in France32.
PCR, transformations and mutant generation
Colony PCRs were performed using OneTaq DNA polymerase (New England Biolabs). Sanger sequencing reactions were prepared using BigDye Terminator Mix v3.1 and sequenced by the Genetic Research Services, UQ. Electrocompetent cells were prepared, and transformations were performed as previously described74. All mutants were constructed using λ-Red-mediated homologous recombination as previously described74. All constructs were confirmed by Sanger sequencing. DNA fragments for cloning and mutations were amplified using KAPA HiFi polymerase (Roche). All primers were synthesised by IDT (Singapore), and a full list of primers used is available in Supplementary Data 3.
Construction of 233 ColVLP database
ColVLP-associated loci from the ColV plasmid pMS7163A32 were used as a BLASTn (v2.14.0+) query against the NCBI RefSeq26 (21,034 chromosomes; 19,736 plasmids; 02/01/2021) and PLSDB25 (34,513 plasmids; 23/06/2021) databases to identify a non-redundant collection of 233 ColVLPs as described in Liu et al.6. The NCBI RefSeq contained the complete genomes of 820 Klebsiella spp., 930 Salmonella spp., and 1441 Escherichia spp. isolates. Incomplete plasmids with ambiguous nucleotides were removed from further analyses. ColVLPs were assigned to pre-defined PTUs using COPLA75, using both linear and circular topologies with default options.
Cladogram and network construction
All unique ORFs from the constituent plasmid sequences were concatenated into a single hypothetical plasmid. The hypothetical plasmid was subsequently compared against each plasmid in a sequence similarity search, generating a binary ORF presence/absence sequence30. All binary sequences were subject to hierarchical clustering based on Manhattan distance using the stats package in the R (v4.3.2) environment76, and visualised as a midpoint-rooted cladogram using the interactive Tree of Life77. Cladogram clusters (>1) were determined using the total sum or square method. Split decomposition networks were generated using the ORF-based binarised structure network analysis of plasmids tool30, and visualised using SplitsTree (v6.3.10)78. Network clusters were determined with a Dice index threshold of >0.71.
General bioinformatic analyses
Pairwise alignments were performed using BLASTn79 and visualised with EasyFig (v2.2.2)80 with a minimum coverage length of 500 bp. Presence/absence of ColVLP-associated genes was determined using pMS7163A sequences32 as a BLASTn query at a 95% alignment length and 90% percent identity threshold. Operons were only considered present if all constituent genes were identified. Conjugative ability was predicted based on the presence of all genes required for F plasmid conjugation81 using a BLASTn search at a 90% alignment length and percent identity threshold. F plasmid replicons were determined using a query against the IncF pMLST scheme28,82 with an exact match threshold. Resistance genes and non-F plasmid replicons were determined using ABRIcate (v0.8.10) (https://github.com/tseemann/abricate) against the ARG-ANNOT83 and PlasmidFinder41 databases, respectively, at a 100% query length threshold. Host E. coli STs were determined using the MLST (v2.23.0) tool (https://github.com/tseemann/mlst) and the PubMLST database82. All features are available in Supplementary Data 4.
Identification of ColVLP co-integrates
ColVLP-associated loci from the ColV plasmid pMS7163A32 were used as a BLASTn query against NCBI (query performed 4th April 2024) as described in Liu et al.6 to identify ColVLPs. Incomplete plasmids with ambiguous nucleotides and partial sequences were removed from further analyses. Plasmid replicon queries from PlasmidFinder41 were used as a BLASTn query against the ColVLPs at a 90% identity and 100% alignment length threshold to identify non-F replicons.
Plasmid carriage inference and plasmid co-carriage network generation
ColVLP and plasmid carriage were inferred using the criteria of Liu et al.6 and PlasmidFinder41, respectively, for the following draft genomes databases: (i) 100ST database consisting of ~100 random strains from the top 100 STs in Enterobase (downloaded on 18th December 2020)49,84; (ii) ST95 database consisting of 2118 strains from Enterobase (downloaded on 24th May 2021)84; (iii) ST131 database consisting of 3857 strains from Enterobase (downloaded 17th July 2018)84, which was additionally separated into their clades as described previously85. All predicted co-carriage events (n > 10) were quantified and plotted as a network using Cytoscape (v3.9.1)86. ColVLPs and IncF plasmids were plotted as separate nodes (i.e., ColVLP positive draft genomes were considered as F plasmid negative).
Virulence gene phylogenetic analyses
Virulence genes were identified from the NCBI RefSeq26 E. coli chromosomal database (1,377 E. coli chromosomes; 02/01/2021) and the 233 ColVLP databases using the loci from pMS7163A32 as a BLASTn query at a 60% alignment length threshold. Nucleotide sequences were aligned using Clustal Omega (v1.2.3)87, built into a phylogenetic tree using FastTree (v2.1)88, and visualised as a midpoint-rooted phylogenetic tree using the interactive Tree of Life77.
OmpT homologue comparison
OmpT prevalence was determined using MS7163 ompT1 (MS7163_RS03130)32 as a query against the NCBI RefSeq database (1377 complete assemblies; 02/01/2021) at a 65% query length and alignment threshold. Predicted protein structures were generated using ColabFold (v1.5.2)89 and annotated using ChimeraX (v1.6.1)90. Amino acid sequence alignments were performed and visualised using CLC Main Workbench (v23.0.4) (QIAGEN).
FRET assay
The synthetic FRET peptide [2Abz-SLGRKIQI-K(Dnp)-NH2] with ortho-aminobenzoic acid (Abz) as the fluorescent and group 2,4 nitrophenyl (Dnp) as the quencher was purchased from Mimotopes at a purity of >95% and resuspended in sterile Milli-Q water. The FRET protocol was adapted from Thomassin et al.91. Briefly, bacterial cells were grown to mid-log phase in NMM and standardised to an OD600 of 0.5 in PBS. Bacterial cells were mixed with the FRET peptide at final concentrations of 3 × 108 CFU/mL and 3 µM (in PBS), respectively, and incubated at 23 °C for 60 min, with fluorescence observation (excitation wavelength 325 nm, emission wavelength 430 nm) every 5 min. Data were normalised to time 0 readouts.
LL-37 Proteolytic cleavage and LC/MS quantification
LL-37 peptides (LLGDFFRKSKEKIGKEFKRIVQRIKDFLRNLVPRTES) were purchased from Mimotopes and resuspended in sterile Milli-Q water. Bacterial cells were grown to mid-log phase in NMM and standardised to OD600 1.8. Bacterial cells were mixed with LL-37 in PBS, at final concentrations of 7 × 107 CFU/mL and 0.1 mg/mL, respectively, and incubated at 37 °C for 2 h. After incubation, trifluoroacetic acid was added to a final concentration of 1%. Bacterial cells were pelleted through centrifugation, and the supernatant was filtered through a 0.22 µm PVDF filter (Merck; SLGV033RS). Samples were stored at −80 °C prior to being analysed by LC-MS on a Shimadzu Nexera, LC30, uHPLC (Japan) coupled to a Triple Tof 5600 mass spectrometer (ABSCIEX, Canada) equipped with a duo electrospray ion source. Samples (20 µL) were injected onto a 2 × 150 mm Phenomenex, Jupiter, 300A, 3.5 µm C18 column (Phenomenex, Australia) at 200 µl/min. Peptides were eluted using linear gradients of 2–60% solvent B over 60 min, followed by 60–80% solvent B over 2 min, and finally 80-98% solvent B in 0.5 min. Solvent A consisted of 0.1% formic acid (aq) and solvent B contained acetonitrile/ 0.1% formic acid (aq). The ionspray voltage was set to 5500 V, declustering potential (DP) 100 V, curtain gas flow 25, nebuliser gas 1 (GS1) 50, GS2 to 60, interface heater at 150 °C and the turbo heater to 500 °C. A 400 ms full scan TOF-MS spectrum over the mass range 500–2000 m/z was acquired and processed using Analyst TF 1.6 software (ABSCIEX, Canada). Quantitation of products corresponding to LL-37AA19-37 was performed using MultiQuant 3.0.3 (ABSCIEX, Canada). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE92 partner repository with the dataset identifier PXD060817.
LL-37 time-kill assay
Bacterial cells were grown in N-minimal medium to mid-log and standardised to an OD600 of 1.8 in PBS. Bacterial cells were mixed with LL-37 (in PBS) at final concentrations of ~4 × 108 CFU/mL cells and 0.05 mg/mL, respectively, and incubated at 37 °C for 2 h. After incubation, bacterial cells were serially diluted, spotted on LB agar, and enumerated the next day. Data is represented as a percentage of the CFU count in the PBS control at time 0.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The LC-MS data generated in this study have been deposited in the ProteomeXchange Consortium via the PRIDE92 partner repository under the accession code PXD060817. The metadata for the 233 ColVLP, 100ST, 1377 NCBI RefSeq complete genomes, ST95 draft genomes, and ST131 Clade B draft genomes databases generated in this study are provided in the Source data. Source data are located in the Supplementary Data files. Sequence data used to verify plasmid constructs is available from the authors upon request. All other data generated in this work are presented in the manuscript.
References
Shintani, M., Sanchez, Z. K. & Kimbara, K. Genomics of microbial plasmids: classification and identification based on replication and transfer systems and host taxonomy. Front. Microbiol. 6, 242 (2015).
Pilla, G. & Tang, C. M. Going around in circles: virulence plasmids in enteric pathogens. Nat. Rev. Microbiol. 16, 484–495 (2018).
Waters, V. L. & Crosa, J. H. Colicin V virulence plasmids. Microbiol. Rev. 55, 437–450 (1991).
Johnson, T. J., Siek, K. E., Johnson, S. J. & Nolan, L. K. DNA sequence of a ColV plasmid and prevalence of selected plasmid-encoded virulence genes among avian Escherichia coli strains. J. Bacteriol. 188, 745–758 (2006).
Johnson, T. J. et al. Identification of minimal predictors of avian pathogenic Escherichia coli virulence for use as a rapid diagnostic tool. J. Clin. Microbiol. 46, 3987–3996 (2008).
Liu, C. M. et al. Escherichia coli ST131-H22 as a foodborne uropathogen. mBio 9, e00470-18 (2018).
Cummins, M. L., Reid, C. J. & Djordjevic, S. P. F plasmid lineages in Escherichia coli ST95: implications for host range, antibiotic resistance, and zoonoses. mSystems 7, e01212-21 (2022).
Reid, C. J., Cummins, M. L. & Djordjevic, S. P. Major F plasmid clusters are linked with ColV and pUTI89-like marker genes in bloodstream isolates of Escherichia coli. BMC Genom. 26, 57 (2025).
Reid, C. J. et al. A role for ColV plasmids in the evolution of pathogenic Escherichia coli ST58. Nat. Commun. 13, 683 (2022).
Fath, M. J., Skvirsky, R., Gilson, L., Mahanty, H. K. & Kolter, R. The secretion of Colicin V. In Bacteriocins, Microcins and Lantibiotics (eds James, R., Lazdunski, C. & Pattus F.) 331–348 (Springer Berlin Heidelberg, 1992).
Cascales, E. et al. Colicin biology. Microbiol. Mol. Biol. Rev. 71, 158–229 (2007).
Frick, K. K., Quackenbush, R. L. & Konisky, J. Cloning of immunity and structural genes for colicin V. J. Bacteriol. 148, 498–507 (1981).
Milch, H., Nikolnikov, S. & Czirók, E. Escherichia coli Col V plasmids and their role in pathogenicity. Acta Microbiol. Hung. 31, 117–125 (1984).
Chagneau, C. V. et al. HlyF, an underestimated virulence factor of uropathogenic Escherichia coli. Clin. Microbiol. Infect. 29, 1449.e1–1449.e9 (2023).
Murase, K. et al. HlyF produced by extraintestinal pathogenic Escherichia coli is a virulence factor that regulates outer membrane vesicle biogenesis. J. Infect. Dis. 213, 856–865 (2016).
Johnson, T. J., Johnson, S. J. & Nolan, L. K. Complete DNA sequence of a ColBM plasmid from avian pathogenic Escherichia coli suggests that it evolved from closely related ColV virulence plasmids. J. Bacteriol. 188, 5975–5983 (2006).
Peigne, C. et al. The plasmid of Escherichia coli strain S88 (O45:K1:H7) that causes neonatal meningitis is closely related to avian pathogenic E. coli plasmids and is associated with high-level bacteremia in a neonatal rat meningitis model. Infect. Immun. 77, 2272–2284 (2009).
Biggel, M. et al. Horizontally acquired papGII-containing pathogenicity islands underlie the emergence of invasive uropathogenic Escherichia coli lineages. Nat. Commun. 11, 5968 (2020).
Desloges, I. et al. Identification and characterization of OmpT-like proteases in uropathogenic Escherichia coli clinical isolates. MicrobiologyOpen 8, e915 (2019).
Khajanchi, B. K. et al. Comparative genomic analysis and characterization of incompatibility group FIB plasmid encoded virulence factors of Salmonella enterica isolated from food sources. BMC Genom. 18, 570 (2017).
Johnson, T. J. et al. Associations between multidrug resistance, plasmid content, and virulence potential among extraintestinal pathogenic and commensal Escherichia coli from humans and poultry. Foodborne Pathog. Dis. 9, 37–46 (2012).
Moran, R. A. & Hall, R. M. Evolution of regions containing antibiotic resistance genes in FII-2-FIB-1 ColV-Colla virulence plasmids. Microb. Drug Resist. 24, 411–421 (2017).
Kai-Larsen, Y. et al. Uropathogenic Escherichia coli modulates immune responses and its curli fimbriae interact with the antimicrobial peptide LL-37. PLoS Pathog. 6, e1001010 (2010).
Chromek, M. et al. The antimicrobial peptide cathelicidin protects the urinary tract against invasive bacterial infection. Nat. Med. 12, 636–641 (2006).
Schmartz, G. P. et al. PLSDB: advancing a comprehensive database of bacterial plasmids. Nucleic Acids Res. 50, D273–D278 (2021).
O’Leary, N. A. et al. Reference sequence (RefSeq) database at NCBI: current status, taxonomic expansion, and functional annotation. Nucleic Acids Res. 44, D733–D745 (2016).
Wang, Y. & Dagan, T. The evolution of antibiotic resistance islands occurs within the framework of plasmid lineages. Nat. Commun. 15, 4555 (2024).
Villa, L., García-Fernández, A., Fortini, D. & Carattoli, A. Replicon sequence typing of IncF plasmids carrying virulence and resistance determinants. J. Antimicrob. Chemother. 65, 2518–2529 (2010).
Moran, R. A., Holt, K. E. & Hall, R. M. pCERC3 from a commensal ST95 Escherichia coli: a ColV virulence-multiresistance plasmid carrying a sul3-associated class 1 integron. Plasmid 84-85, 11–19 (2016).
Suzuki, M., Doi, Y. & Arakawa, Y. ORF-based binarized structure network analysis of plasmids (OSNAp), a novel approach to core gene-independent plasmid phylogeny. Plasmid 108, 102477 (2020).
Lian, Z. J. et al. Genetic basis of I-complex plasmid stability and conjugation. PLOS Genet. 19, e1010773 (2023).
Nhu, N. T. K. et al. Discovery of new genes involved in curli production by a uropathogenic Escherichia coli strain from the highly virulent O45:K1:H7 lineage. mBio 9, e01462-18 (2018).
Wyrsch, E. R. et al. Complete sequences of multiple-drug resistant IncHI2 ST3 plasmids in Escherichia coli of porcine origin in Australia. Front. Sustain. Food Syst. 3, 18 (2019).
Hall, R. M. & Vockler, C. The region of the IncN plasmid R46 coding for resistance to beta-lactam antibiotics, streptomycin/spectinomycin and sulphonamides is closely related to antibiotic resistance segments found in IncW plasmids and in Tn21-like transposons. Nucleic Acids Res. 15, 7491–7501 (1987).
Harmer, C. J., Moran, R. A. & Hall, R. M. Movement of IS26-associated antibiotic resistance genes occurs via a translocatable unit that includes a single IS26 and preferentially inserts adjacent to another IS26. mBio 5, e01801-14 (2014).
Harmer, C. J. & Hall, R. M. IS26 and the IS26 family: versatile resistance gene movers and genome reorganizers. Microbiol. Mol. Biol. Rev. 88, e0011922 (2024).
McKinnon, J., Roy Chowdhury, P. & Djordjevic, S. P. Molecular analysis of an IncF ColV-like plasmid lineage that carries a complex resistance locus with a trackable genetic signature. Microb. Drug Resist. 26, 787–793 (2020).
Sun, Y. W. et al. IS26-flanked composite transposon Tn6539 carrying the tet(M) gene in IncHI2-type conjugative plasmids from Escherichia coli isolated from ducks in China. Front. Microbiol. 9, 3168 (2018).
Harmer, C. J. & Hall, R. M. Targeted conservative cointegrate formation mediated by IS26 family members requires sequence identity at the reacting end. mSphere 6, e01321-20 (2021).
Harmer, C. J. & Hall, R. M. Targeted conservative formation of cointegrates between two DNA molecules containing IS26 occurs via strand exchange at either IS end. Mol. Microbiol. 106, 409–418 (2017).
Carattoli, A. et al. In silico detection and typing of plasmids using PlasmidFinder and plasmid multilocus sequence typing. Antimicrob. Agents Chemother. 58, 3895–3903 (2014).
Seenama, C., Thamlikitkul, V. & Ratthawongjirakul, P. Multilocus sequence typing and blaESBL characterization of extended-spectrum beta-lactamase-producing Escherichia coli isolated from healthy humans and swine in Northern Thailand. Infect. Drug Resist. 12, 2201–2214 (2019).
Emery, A., Hocquet, D., Bonnet, R. & Bertrand, X. Genotypic characteristics and antimicrobial resistance of Escherichia coli ST141 clonal group. Antibiotics 12, 382 (2023).
Ćwiek, K. et al. Phenotypic and genotypic characterization of mcr-1-positive multidrug-resistant Escherichia coli ST93, ST117, ST156, ST10, and ST744 isolated from poultry in Poland. Braz. J. Microbiol. 52, 1597–1609 (2021).
Johnson, T. J. et al. Refining the definition of the avian pathogenic Escherichia coli (APEC) pathotype through inclusion of high-risk clonal groups. Poult. Sci. 101, 102009 (2022).
Nesporova, K. et al. Escherichia coli sequence type 457 is an emerging extended-spectrum-β-lactam-resistant lineage with reservoirs in wildlife and food-producing animals. Antimicrob. Agents Chemother. 65, e01118-20 (2020).
Zając, M. et al. Occurrence and characterization of mcr-1-positive Escherichia coli isolated from food-producing animals in Poland, 2011-2016. Front. Microbiol. 10, 1753 (2019).
Lalaoui, R. et al. Detection of plasmid-mediated colistin resistance, mcr-1 gene, in Escherichia coli isolated from high-risk patients with acute leukemia in Spain. J. Infect. Chemother. 25, 605–609 (2019).
Phan, M. D. et al. Plasmid-mediated ciprofloxacin resistance imparts a selective advantage on Escherichia coli ST131. Antimicrob. Agents Chemother. 66, e0214621 (2022).
Welch, R. A. et al. Extensive mosaic structure revealed by the complete genome sequence of uropathogenic Escherichia coli. Proc. Natl. Acad. Sci. USA 99, 17020–17024 (2002).
Dobrindt, U. et al. Genetic structure and distribution of four pathogenicity islands (PAI I(536) to PAI IV(536)) of uropathogenic Escherichia coli strain 536. Infect. Immun. 70, 6365–6372 (2002).
Dekker, N., Cox, R. C., Kramer, R. A. & Egmond, M. R. Substrate specificity of the integral membrane protease OmpT determined by spatially addressed peptide libraries. Biochemistry 40, 1694–1701 (2001).
McCarter, J. D. et al. Substrate specificity of the Escherichia coli outer membrane protease OmpT. J. Bacteriol. 186, 5919–5925 (2004).
Ridyard, K. E. & Overhage, J. The potential of human peptide LL-37 as an antimicrobial and anti-biofilm agent. Antibiotics 10, 650 (2021).
Manges, A. R. et al. Global extraintestinal pathogenic Escherichia coli (ExPEC) lineages. Clin. Microbiol. Rev. 32, e00135-18 (2019).
Xia, F. et al. Complete genomic analysis of ST117 lineage extraintestinal pathogenic Escherichia coli (ExPEC) to reveal multiple genetic determinants to drive its global transmission: ST117 E. coli as an emerging multidrug-resistant foodborne ExPEC with zoonotic potential. Transbound. Emerg. Dis. 69, 3256–3273 (2022).
Santos, A. C. M. et al. Virulence potential of a multidrug-resistant Escherichia coli strain belonging to the emerging clonal group ST101-B1 isolated from bloodstream infection. Microorganisms 8, 827 (2020).
Yamaji, R. et al. A population-based surveillance study of shared genotypes of Escherichia coli isolates from retail meat and suspected cases of urinary tract infections. mSphere 3, e00179-18 (2018).
Li, Z. et al. Antimicrobial resistance and genomic characteristics of Escherichia coli strains isolated from the poultry industry in Henan Province, China. Microorganisms 12, 575 (2024).
Marr, C. M. & Russo, T. A. Hypervirulent Klebsiella pneumoniae: a new public health threat. Expert Rev. Anti-infect. Ther. 17, 71–73 (2019).
Lam, M. M. C. et al. Tracking key virulence loci encoding aerobactin and salmochelin siderophore synthesis in Klebsiella pneumoniae. Genome Med. 10, 77 (2018).
Russo, T. A. et al. Aerobactin mediates virulence and accounts for increased siderophore production under iron-limiting conditions by hypervirulent (hypermucoviscous) Klebsiella pneumoniae. Infect. Immun. 82, 2356–2367 (2014).
Shankar, C. et al.Emergence of multidrug resistant hypervirulent ST23 Klebsiella pneumoniae: multidrug resistant plasmid acquisition drives evolution. Front. Cell. Infect. Microbiol. 10, 575289 (2020).
Dong, N., Lin, D., Zhang, R., Chan, E. W.-C. & Chen, S. Carriage of blaKPC-2 by a virulence plasmid in hypervirulent Klebsiella pneumoniae. J. Antimicrob. Chemother. 73, 3317–3321 (2018).
Jin, L., Wang, R., Gao, H., Wang, Q. & Wang, H. Identification of a novel hybrid plasmid encoding KPC-2 and virulence factors in Klebsiella pneumoniae sequence type 11. Antimicrob. Agents and Chemother. 65, https://doi.org/10.1128/aac.02435-20 (2021).
Dionisio, F., Zilhão, R. & Gama, J. A. Interactions between plasmids and other mobile genetic elements affect their transmission and persistence. Plasmid 102, 29–36 (2019).
Liu, Y. Y. et al. The formation of two hybrid plasmids mediated by IS26 and Tn6952 in Salmonella enterica serotype enteritidis. Front. Microbiol. 12, 676574 (2021).
Goze, A. & Ehrlich, S. D. Replication of plasmids from Staphylococcus aureus in Escherichia coli. Proc. Natl. Acad. Sci. USA 77, 7333–7337 (1980).
Woodward, M. J., Wray, C., Ridha, G. A. & Walton, J. R. Plasmid and chromosomal related toxin polymorphism of Escherichia coli serogroup O 138; plasmid transfer and co-integration with pRP4. J. Med. Microbiol. 31, 241–249 (1990).
Froehlich, B., Parkhill, J., Sanders, M., Quail, M. A. & Scott, J. R. The pCoo plasmid of enterotoxigenic Escherichia coli is a mosaic cointegrate. J. Bacteriol. 187, 6509–6516 (2005).
Brannon, J. R., Thomassin, J. L., Desloges, I., Gruenheid, S. & Le Moual, H. Role of uropathogenic Escherichia coli OmpT in the resistance against human cathelicidin LL-37. FEMS Microbiol. Lett. 345, 64–71 (2013).
Kramer, R. A. et al. Lipopolysaccharide regions involved in the activation of Escherichia coli outer membrane protease OmpT. Eur. J. Biochem. 269, 1746–1752 (2002).
Sinsinbar, G. et al. Role of lipopolysaccharide in protecting OmpT from autoproteolysis during in vitro refolding. Biomolecules 10, 922 (2020).
Kakkanat, A. et al. The role of H4 flagella in Escherichia coli ST131 virulence. Sci. Rep. 5, 16149 (2015).
Redondo-Salvo, S. et al. COPLA, a taxonomic classifier of plasmids. BMC Bioinform. 22, 390 (2021).
Team RC. 2023. R: a language and environment for statistical computing, https://www.R-project.org/.
Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res. https://doi.org/10.1093/nar/gkae268 (2024).
Huson, D. H. & Bryant, D. Application of phylogenetic networks in evolutionary studies. Mol. Biol. Evol. 23, 254–267 (2006).
Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinform. 10, 421 (2009).
Sullivan, M. J., Petty, N. K. & Beatson, S. A. Easyfig: a genome comparison visualizer. Bioinformatics 27, 1009–1010 (2011).
Fernandez-Lopez, R., de Toro, M., Moncalian, G., Garcillan-Barcia, M. P. & de la Cruz, F. Comparative genomics of the conjugation region of F-like plasmids: five shades of F. Front. Mol. Biosci. 3, 71 (2016).
Jolley, K. A., Bray, J. E. & Maiden, M. C. J. Open-access bacterial population genomics: BIGSdb software, the PubMLST.org website and their applications. Wellcome Open Res. 3, 124 (2018).
Gupta, S. K. et al. ARG-ANNOT, a new bioinformatic tool to discover antibiotic resistance genes in bacterial genomes. Antimicrob. Agents Chemother. 58, 212–220 (2014).
Dyer, N. igelP. et al. EnteroBase in 2025: exploring the genomic epidemiology of bacterial pathogens. Nucleic Acids Res. 53, D757–D762 (2024).
Petty, N. K. et al. Global dissemination of a multidrug resistant Escherichia coli clone. Proc. Natl. Acad. Sci. USA 111, 5694–5699 (2014).
Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504 (2003).
Sievers, F. & Higgins, D. G. Clustal Omega for making accurate alignments of many protein sequences. Protein Sci. 27, 135–145 (2018).
Price, M. N., Dehal, P. S. & Arkin, A. P. FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Mol. Biol. Evol. 26, 1641–1650 (2009).
Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat. Methods 19, 679–682 (2022).
Pettersen, E. F. et al. UCSF ChimeraX: Structure visualization for researchers, educators, and developers. Protein Sci. 30, 70–82 (2021).
Thomassin, J. L., Brannon, J. R., Gibbs, B. F., Gruenheid, S. & Le Moual, H. OmpT outer membrane proteases of enterohemorrhagic and enteropathogenic Escherichia coli contribute differently to the degradation of human LL-37. Infect. Immun. 80, 483–492 (2012).
Perez-Riverol, Y. et al. The PRIDE database at 20 years: 2025 update. Nucleic Acids Res. 53, D543–D553 (2025).
Acknowledgements
The authors thank the University of Queensland Institute for Molecular Bioscience colleagues Nicholas Yuen for providing technical assistance and Alun Jones for advice and assistance with LC/MS quantification of the LL-37 cleavage products. This research was supported by use of the Nectar Research Cloud and by QCIF. The Nectar Research Cloud is a collaborative Australian research platform supported by the NCRIS-funded Australian Research Data Commons (ARDC). The work was supported by grants from the Australian National Health and Medical Research Council (APP2001431 to M.A.S., M.-D.P. and N.T.K.N.; APP2037698 to M.A.S., M.-D.P.), the Australian Medical Research Future Fund (MRFGH000028 to M.A.S., M.-D.P.) and the Australian Research Council (DP230101930 to M.A.S.).
Author information
Authors and Affiliations
Contributions
Conceptualisation: Z.J.L., N.T.K.N., M.-D.P., M.A.S. Investigation: Z.J.L., N.T.K.N., C.R., C.C., I.M.-R. Formal analysis: Z.J.L., N.T.K.N., M.-D.P., M.A.S. Supervision: M.-D.P., M.A.S. Writing—original draft: Z.J.L., M.-D.P., M.A.S. Writing—review and editing: all authors. All authors read and approved the final manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Nabil-Fareed Alikhan and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Lian, Z.J., Nhu, N.T.K., Ravi, C. et al. Convergence of plasmid-driven virulence and antibiotic resistance in Escherichia coli. Nat Commun 17, 505 (2026). https://doi.org/10.1038/s41467-025-67202-9
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41467-025-67202-9
