The E. albertii strain ESA311, assigned to sequence type ST4606 and serotype EAOg1:EAHg4, was isolated from an infant suffering from diarrhea in Sichuan province, China (Zhang et al., 2025). E. coli EC600 (rifampin-resistant) was used as the recipient strain in conjugation experiments. Bacteria were routinely grown at 37°C on either solid or liquid Luria-Bertani (LB) medium (Beijing Land Bridge Technology Co., Ltd., Beijing, China).

The antimicrobial susceptibility profile was evaluated using the NMIC/ID-801 Panels on the BD Phoenix™ 100 automated system (Becton Dickinson Company, Sparks, MD, United States). The panel included 21 antibiotics across 9 drug classes: phenicols (chloramphenicol), β-lactams (ampicillin, cefotaxime, ceftazidime, cefoxitin, aztreonam, ceftazidime-avibactam, ampicillin-sulbactam, ertapenem, meropenem, imipenem), tetracyclines (tetracycline, tigecycline), macrolides (azithromycin), quinolones (ciprofloxacin, nalidixic acid), aminoglycosides (amikacin, streptomycin), diaminopyrimidines and sulfonamides (trimethoprim-sulfamethoxazole), nitrofurans (nitrofurantoin), and polymyxins (colistin). E. coli ATCC 25922 was used for quality control. Antimicrobial susceptibility was interpreted according to the Clinical and Laboratory Standards Institute (CLSI M100 ED34:2024) for Enterobacteriaceae (CLSI, 2024).

Genomic DNA of the E. albertii strain ESA311 was extracted from an overnight fresh culture using the Wizard^® Genomic DNA purification kit (Promega, Madison, WI, United States), and subsequently sequenced on both the PacBio Sequel II and DNBSEQ platforms at the Beijing Genomics Institute (BGI, Shenzhen, China). Detailed sequencing and assembly procedures have been described in our previous study (Zhang et al., 2025). Coding sequences (CDSs) and pseudogenes were predicted using Prokka v1.14.6 (Seemann, 2014) and RAST (Accessed 20 December 2024),¹ and further annotated through BLASTP and BLASTN searches against the UniProtKB/Swiss-Prot (Accessed 24 December 2024)² and NCBI non-redundant (NR) database (Sayers et al., 2024).

The presence of ARGs in the genome of ESA311 was determined using ABRicate 1.0.1³ with default parameters against the ResFinder database (Zankari et al., 2012). The insertion sequences (ISs) flanking the ARGs in the genome of strain ESA311 were identified using ISfinder (Siguier et al., 2006). The Integrons containing the integrase (intI) and cassettes encoding accessory genes were detected and analyzed using IntegronFinder 2.0 (Néron et al., 2022). Additionally, potential virulence factors were screened by querying the Virulence Factor Database (VFDB) under default settings (Chen et al., 2016).

The replicon types of the plasmids contained by strain ESA311 were determined using PlasmidFinder 2.1.⁴ The analysis was conducted employing the database “Enterobacteriaceae”. Unless otherwise specified, the analysis was performed with a minimum coverage of 60% and a minimum identity of 95%.

Conjugation experiments were carried out using a filter mating method, with ESA311 as the donor strain and E. coli EC600 (rifampin-resistant) as the recipient strain (Yuan et al., 2021). Briefly, both strains were cultured in LB broth and incubated overnight at 37°C. For each conjugation assay, 200 μL of donor culture were mixed with 600 μL of recipient culture (v:v = 1:3). The mixture was vortexed thoroughly and 100 μL of the suspension was transferred onto a cellulose filter membrane (pore size, 0.22 μm) placed on a LB agar plate. Following incubation at 37 °C for 16–18 h, bacteria from the membrane were scraped off and resuspended in 1 mL of sterile saline solution. Serial dilutions of the resuspension were prepared and plated onto selective LB agar media. Transconjugants carrying plasmid pESA311_1 were selected on LB agar plates supplemented with 50 μg/mL chloramphenicol and 100 μg/mL rifampin, whereas those carrying pESA311_2 were selected using 2 μg/mL ciprofloxacin and 100 μg/mL rifampin. Transfer efficiency was calculated based on colony counts of transconjugants relative to recipient cells, performed in triplicate (Zhu et al., 2013). To confirm the presence of the respective plasmids in the transconjugants, eight randomly selected colonies from each selection plate were subjected to PCR analysis using plasmid-specific primers (Supplementary Table 1).

To construct phylogenetic trees, we identified plasmid sequences from NCBI that shared a minimum identity of 95% with pESA311_1 (≥ 80% coverage; n = 114) and pESA311_2 (≥ 60% coverage; n = 20), as detailed in Supplementary Tables 2, 3, respectively. QUAST v5.2.0 was used to assess the quality of the downloaded genomes (Mikheenko et al., 2018). A pairwise Mash distance matrix was calculated using Mash v2.3 (Ondov et al., 2016). This matrix was then used to infer a phylogenetic tree via the Neighbor-Joining (NJ) method as implemented in FastME v2.1.6.3 (Lefort et al., 2015).

A chi-square test was conducted to evaluate the association between the presence of intI1 and MDR. Statistical analysis was performed using IBM SPSS Statistics 21, with a p < 0.05 considered statistically significant.

Circular plasmid maps were generated using Proksee (Grant et al., 2023) and subsequently annotated for clarity. Comparisons of MDR regions and generation of linear genomic maps were conducted using Easyfig v2.2.5 (Sullivan et al., 2011). All schematic diagrams were assembled and finalized using vector graphics software (Inkscape v1.4 or Adobe Illustrator 2020). The butterfly diagram and chord diagram were generated using the online data visualization platform at https://www.cnsknowall.com/ (accessed on 12 March 2025).

The nucleotide sequence of the ESA311 chromosome has been deposited in GenBank under the accession number CP157783. The complete sequences of plasmids pESA311_1 through pESA311_5 have been submitted to GenBank under accession numbers CP157784 to CP157788. The accession numbers for the other plasmids and strains used in this study are provided in Supplementary Tables 2–4, respectively.

Initial antimicrobial susceptibility testing performed at the clinical microbiology laboratory using the Siemens LAB POR system indicated that strain ESA311 was susceptible to cefotaxime (MIC ≤ 2 μg/mL) and other third-generation cephalosporins (Zhang et al., 2025). However, therapeutic failure following cefotaxime monotherapy in the infected neonate prompted comprehensive reevaluation of its resistance profile. To resolve this discrepancy, confirmatory antimicrobial susceptibility testing was performed using the NMIC/ID-801 panels on the BD Phoenix™ 100 automated system. This advanced platform revealed several critical reclassifications in susceptibility results (Table 1). Notably, Retesting results revealed an elevated minimal inhibitory concentration (MIC) for cefotaxime (16 μg/mL), classifying the strain as resistant—contrasting with the initial clinical report of susceptibility (≤ 2 μg/mL)—and aligning with the observed clinical treatment failure.

Importantly, strain ESA311 remained susceptible to carbapenems, with a meropenem MIC of 0.25 μg/mL, suggesting potential alternative therapeutic options. However, ESA311 exhibited resistance to multiple non-β-lactam antibiotics including chloramphenicol, tetracycline, azithromycin, ciprofloxacin, nalidixic acid, and trimethoprim-sulfamethoxazole. Notably, it retained susceptibility to certain aminoglycosides such as amikacin and streptomycin, highlighting the complexity of its resistance profile.

Whole-genome sequencing (WGS) revealed that the genome of E. albertii strain ESA311 consists of a single chromosome of 4,795,532 bp and five plasmids. The average GC content of the chromosome is 49.8%. All five plasmids have circular closed DNA sequences, designated pESA311_1 to pESA311_5, with sizes ranging from 42,365 to 209,400 bp. Plasmid characteristics, including size, GC content, and replicon types, are summarized in Table 2.

Plasmid pESA311_1 (209,400 bp) contains 276 predicted CDSs and has an overall GC content of 47.4% (Figure 1A). It harbors two replicon types: IncHI2, showing 100% coverage and 100% sequence identity to the IncHI2 reference sequence of plasmid R478 (BX664015), and IncHI2A, which also exhibits 100% coverage but 99.52% sequence identity to the IncHI2A reference sequence of plasmid R478 (BX664015). Plasmid pESA311_2 (95,634 bp), containing 130 CDSs and a GC content of 54.4% (Figure 1B), was classified as an IncFII-type plasmid based on its alignment to the IncFII (pHN7A8) reference sequence (JN232517), exhibiting 100% coverage and 96.18% sequence identity to the replicon region. Similarly, pESA311_3 (89,218 bp, 121 CDSs, 50% GC content) (Figure 1C) displayed 100% coverage and 100% sequence identity when aligned to the IncI1α reference replicon sequence (AP005147). In contrast, pESA311_4 (51,081 bp, 91 CDSs, 45.4% GC content) (Figure 1D) could not be initially assigned to a specific replicon type. However, by lowering the identity threshold to 90%, we identified two replicon types—IncFIB (reference sequence AP001918) and IncFII (pSE11; reference sequence AP009242)—both with 100% coverage and identities of 93.84 and 93.94%, respectively. Finally, pESA311_5 (42,365 bp, 54 CDSs, 45.1% GC content) (Figure 1E) was identified as an IncP-1 plasmid based on its alignment to reference sequence CM007914.

Genomic annotation revealed distinct conjugative gene architectures among the five plasmids. pESA311_1 carries a set of tra genes (e.g., traC, traV, traB, traK, traE, traL). pESA311_2 possesses a complete IncF-type conjugative system, encompassing genes from traA to traX, including the relaxase gene traI and coupling protein gene traD (Figures 1A,B). pESA311_3 encodes a full IncI-type conjugative module with a comprehensive tra/trb cluster and the relaxase gene nikB. pESA311_4 harbors a truncated set of IncF-like tra genes (traI, traA, traE, traL, traY, traJ, finO). pESA311_5 contains key IncP-type transfer genes, including relaxase gene traI, coupling protein gene traG, and trb genes. Notably, pESA311_4 and pESA311_5 lack several core components typically found in autonomous conjugative systems, such as complete type four secretion system (T4SS) gene clusters for pilus assembly (Lawley et al., 2003; Fronzes et al., 2009).

Searches against the ResFinder database revealed the presence of multiple ARGs on plasmids pESA311_1 and pESA311_2 (Table 1). In contrast, only a single ARG, the major facilitator superfamily (MFS) efflux pump gene mdf(A), was identified on the chromosome of strain ESA311. This gene confers resistance to quinolones and contributes to MDR (Swick et al., 2011).

The MDR plasmids pESA311_1 and pESA311_2 harbor a diverse repertoire of ARGs, conferring resistance to multiple classes of antibiotics, including phenicols (cmlA, floR), β-lactams (bla_OXA–1, bla_TEM–1B), tetracyclines [tet(A)], macrolides [mef(B), mph(A)], quinolones (qnrS1), aminoglycosides (aac(6’)-Ib-cr, aph(4)-Ia, aac(3)- Ia, aac(3)- IId, aadA2, aadA5, ant(3”)- Ia), sulfonamides (sul1, sul2, sul3), and diaminopyrimidines (dfrA12, dfrA17). The detailed visualization of the genetic organization and distribution of these ARGs were presented in Figures 1, 2, illustrating their clustering within specific regions of the plasmids. Such arrangements may facilitate co-selection and horizontal gene transfer.

The other three plasmids lacked detectable ARGs. However, pESA311_4 carried several virulence-associated genes, including virK, espL2, papD, mrkD, papC, and mrkA (Figure 1D). In contrast, pESA311_3 (Figure 1C) and pESA311_5 (Figure 1D) did not harbor any known virulence genes, highlighting the distinct genetic profiles of these plasmids. As shown in Figure 1, detailed annotation of the plasmids revealed differences in their genetic organization and functional potential, with pESA311_4 harboring a unique cluster of virulence genes that may contribute to its pathogenic potential.

To better understand the genetic organization of ARGs in plasmids pESA311_1 and pESA311_2, we analyzed the ARGs and their associated mobile genetic elements. The ARGs and the adjacent MGEs carried by pESA311_1 were clustered within two large MDR regions: a 20,362 bp fragment designated MDR1 and a 32,609 bp fragment designated MDR2 (Figure 1A).

The pESA311_1 MDR1 (Figure 2A) consists of two different transposable units. The first segment corresponds to a Tn1525-like transposon harboring a class I integron (intI1) cassette array [intI1-dfrA12-aadA2-cmlA1-ant(3”)-Ia-qacE-MULE (mutator-like element) —sul3] and a mef(B) gene. The second segment contains an incomplete TnAS1 remnant (ΔTnAS1), followed by the tet(R) and tet(A) genes, and ends with another ΔTnAS1. These two segments are connected by IS26.

The pESA311_1 MDR2 (Figure 2B) is flanked by IS26 and ΔIS692 and comprises a complex arrangement of mobile elements. It begins with an IS26-linked composite transposon containing sul1-qacE-aar3-bla_OXA–1-aac(6’)-Ib-cr, followed by a second IS26-linked composite transposon harboring aac(3’)-IVa-aph(4)-Ia-ISEc59-ΔTn5393-ISAba1- sul2-floR-IS1006-ΔTnAS3. Additional ISs, including IS15DI, IS3H, ΔIS2, and ΔIS692, are also present in this region. Notably, mercury resistance-related genes embedded within an IS15DI element suggest that strain ESA311 may possess genetic determinants for mercury resistance (Figure 1A).

The ARGs and adjacent ISs carried by pESA311_2 were clustered within a 28,830 bp fragment (Figure 2C), which was organized sequentially as a β-lactam resistant gene, bla_TEM–1B, flanked by IS15DI and ΔTn2, IS15DI-mph(R)-mrxA-mph(A)- IS26-tmrB-like-aac(3)-IId module, and a intI1 cassette array (intI1-dfrA17-aadA5-qacE-sul1) carried by IS15DI on the negative strand, with further elements including ΔIS2, qnrS1, ISKpn19, tet(R), tet(A), and ΔTnAS1.

To evaluate the conjugative transfer potential of the antibiotic resistance plasmids pESA311_1 and pESA311_2, conjugation assays were performed using E. coli EC600, a rifampin-resistant strain, as the recipient. Transconjugants were selected on media containing chloramphenicol (50 μg/mL; selecting for pESA311_1) or ciprofloxacin (2 μg/mL; selecting for pESA311_2). The calculated conjugation frequencies were 5.30 × 10⁻³ and 4.90 × 10⁻³, respectively.

To confirm the presence of the plasmids in the transconjugants, eight randomly selected colonies from each antibiotic selection plate were subjected to PCR analysis using plasmid-specific primers (Supplementary Table 1). A striking pattern emerged: all transconjugants obtained under either selection regimen were found to carry a conserved set of plasmids—pESA311_1, pESA311_2, and pESA311_5 (Supplementary Figure 1). Notably, one ciprofloxacin-resistant transconjugant (Transconjugant 1) was found to additionally harbor pESA311_3, while another ciprofloxacin-resistant transconjugant (Transconjugant 7) carried pESA311_4. These observations highlight the potential for co-transfer of multiple plasmids, which may facilitate the spread of MDR among bacterial populations.

The phylogenetic analysis revealed that the pESA311_1 related plasmid sequences were isolated from 1984 to 2023, with a significant increase in prevalence after 2018 (Supplementary Table 2 and Figure 3A). They were primarily hosted by Salmonella enterica and E. coli and exhibited a pronounced geographical clustering, with the overwhelming majority (96.5%) originating from China. Despite sharing the highest global sequence similarity with CP100967.1 (97.0% coverage, 100% identity), pESA311_1 was phylogenetically most closely related to plasmids MT318677.1 and CP051431.1, both of which were obtained from animal-derived strains. The plasmid phylogeny showed a high prevalence of the intI1 integron (111, 89.5%), with ARGs counts ranging from 9 to 27 across these plasmids.

Despite a limited set of 21 homologous plasmids (>60% coverage, >95% identity) identified for pESA311_2, phylogenetic analysis revealed a broad geographical distribution across 12 countries and significant genetic diversity (Supplementary Table 3 and Figure 3B). While the predominant replicon type was IncFII(pHN7A8), matching pESA311_2, other types including IncFII(pSE11), IncFII(pCoo), and IncX1 were also detected. The phylogeny segregated the plasmids into four distinct Clades. Plasmid pESA311_2 resided in Clade II, forming a subclade with its closest relatives, AP041816.1 and CP020338.1. The intI1 integron was universally present in Clades I, II, and III but was absent in Clade IV. Regarding host sources, most isolates with available data were of human origin, with a single strain isolated from the rhesus macaque (Macaca mulatta). Notably, this Macaca mulatta-derived strain lacked any detectable resistance genes, whereas all other strains carried between 3 and 9 resistance genes.

Two distinct intI1-borne gene cassette arrays form integral components of the MDR regions in strain ESA311, with one located on plasmid pESA311_1 and the other on pESA311_2. We analyzed whole-genome sequencing data from 160 clinical E. albertii isolates, comprising publicly available sequences and strains maintained by our research group (Supplementary Table 4). All isolates carried at least one ARG, and 22.5% (36/160) were predicted to be MDR (resistant to three or more tested antibiotic classes as defined by Magiorakos et al., 2012). Our analysis revealed that intI1 was present in 20.0% (32/160) of the isolates.

Notably, the geographical distribution patterns of intI1-positive and multidrug-resistant strains were highly similar (Figure 4A). A strong positive correlation was observed between the presence of intI1 and MDR status (φ = 0.816, p < 0.01). Furthermore, the Chord diagram (Figure 4B) revealed the frequent co-occurrence patterns between intI1 and ARGs conferring resistance to β-lactam antibiotics, tetracycline antibiotics, macrolide antibiotics, and quinolone antibiotics.