A total of 29 unique isolates were included in this study, with each isolate obtained from a different patient. Relevant demographic information such as gender, age, and hospitalized department was retrieved from the electronic data record room. These strains do not represent all E. coli clinical strains isolated during this time but specifically those producing ESBLs with focus on ST410 is due to its clinical significance and contribution to the high ESBL rate observed in our hospital. Specimens were cultured on blood agar and MacConkey’s agar to facilitate bacterial growth. Identification of the isolates was performed using the VITEK^®2 compact system (BioMerieux, France). Confirmatory identification was carried out using the API-20 test and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF/MS) with the Microflex Biotyper^® LT instrument (Bruker Daltonik GmbH, Bremen, Germany) (Patil et al., 2019b). ATCC 29522 E. coli was used as the control strain. The isolates were stored in 40% glycerol at -80°C for further studies.

The antimicrobial susceptibility test (AST) was conducted on all 29 confirmed E. coli isolates using the VITEK^®2 compact system (BioMerieux, France) with the standard AST09 card (software version 9.01). Additionally, the minimum inhibitory concentrations (MICs) for various antimicrobial agents were determined using the broth dilution method. The tested antimicrobial agents included aminoglycosides (amikacin, tobramycin); aminopenicillin (amoxicillin-clavulanate); cephalosporins (2^nd generation- cefuroxime, cefoxitin; 3^rd generation ceftazidime, ceftriaxone, cefoperazone/sulbactam; 4^th generation cefepime; 5^th generation ceftolozane/tazobactam); carbapenems (imipenem, ertapenem); fluoroquinolones (levofloxacin); phosphonic group -(Fosfomycin); nitrofurantoin group (nitrofurantoin); sulfonamide group (trimethoprim/sulfamethoxazole); penicillins plus β-lactamase inhibitors (piperacillin-tazobactam, ticarcillin/clavulanate), and glycylcycline (tigecycline). The AST results were accurately interpreted based on the Clinical and Laboratory Standards Institute (CLSI) (M100Ed34 | Performance Standards for Antimicrobial Susceptibility Testing, 34th Edition, n.d.). The multidrug-resistant (MDR) phenotype was classified according to the criteria defined by Magiorakos et al., where MDR was defined as resistance to at least one agent in three or more antimicrobial groups, and extensively drug-resistant (XDR) was defined as resistance to at least one agent in all but two antimicrobial categories (Magiorakos et al., 2012). The production of ESBL was initially determined using the VITEK^®2 compact system. However, for enhanced accuracy and precision, the results were confirmed using a double-disc synergy test (DDST) and Metallo-beta-lactamase (MBL) production was assessed using the MBL-E-test (IPM-EDTA; AB Biodisk) specifically for ESBL-producing E. coli as per our laboratory’s standardized procedure (Patil et al., 2023a). A positive control strain, which had been previously characterized in our laboratory and confirmed to exhibit the desired traits, was employed as a reference for the experiments (Patil et al., 2019b). In addition, E. coli ATCC 25922, a well-established negative control strain widely used for comparative purposes, was included in the study.

Genomic DNA extraction from the E. coli isolates was carried out using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Subsequently, the Nextera XT DNA Library Prep Kit (Illumina, San Diego, CA, USA) was utilized to prepare the whole-genome sequencing libraries. Multiplexed paired-end sequencing was conducted on an Illumina MiSeq instrument using the MiSeq Reagent V3 Kit (2×300 cycles). To assemble the sequence reads, SPAdes 3.9 software on the ARIES Galaxy server (https://w3.iss.it/site/aries/, accessed on April 14, 2023) was employed for de novo assembly. MLST analysis was performed based on the E. coli MLST website scheme (https://enterobase.warwick.ac.uk/species/index/ecoli, last accessed on April 18, 2023). To analyze E. coli ST410 isolated from a children’s hospital globally, we collected data from the Sequence Read Archive (SRA) at NCBI (https://www.ncbi.nlm.nih.gov/sra/) spanning the years 2009 to 2021 ( Supplementary Table 1 ). Our analysis was based on reports and databases that identified the sequences as ST410. Our objective was to construct a comprehensive global distribution map for E. coli ST410, focusing exclusively on human-origin isolates and excluding isolates from other sources. In silico analysis of the assembled contigs was performed utilizing the available tools at the CGE server. Serotype Finder was employed to predict the serotype of the isolates. Plasmid Finder and ResFinder tools were used to screen the contigs for plasmid and resistance gene content, respectively, at the CGE server. For plasmid subtyping, the replicon allele at the plasmid MLST site was assigned, and the plasmid types were identified using the plasmid MLST database (https://pubmlst.org/plasmid/, accessed on April 21, 2023). Virulence gene identification related to ESBL-E. coli was performed using VirulenceFinder at CGE. Phylogenetic group analysis was performed following the Clermont method (Clermont et al., 2013), which utilizes a triplex PCR approach to classify E. coli into phylogenetic groups A, B1, B2, and D. The specific genetic markers chuA, yjaA, and the DNA fragment TspE4.C2 were targeted to assign the isolates to the respective groups. This method was chosen due to its accuracy and wide acceptance in the scientific community. The detailed protocol is referenced in Patil et al. (2019).

We conducted an experiment to investigate the transfer of resistance genes through broth mating, focusing on the genetic environment of (ESBLs)-encoding-cephalosporin resistance genes. Plasmid transfer frequency was calculated. We used streptomycin-resistant E. coli C₆₀₀ as the recipient strain. Donor and recipient strains were mixed in a 1:1 ratio and incubated at 37°C for 18-24 hours. Transconjugants were selected on Muller–Hinton agar containing 4µg/ml cefotaxime. To confirm the ESBL phenotype in the transconjugants, an E-test was performed followed by PCR assay and sequencing. The primers and protocol were adopted from our lab as outlined in our previous work (Patil et al., 2019a).

The antibiotic susceptibility profile of the clinical isolates was evaluated for commonly used antibiotics in clinical practice. The observed resistance patterns among the isolates indicated substantial rates of resistance, particularly to cephalosporins, followed by aminopenicillin, fluoroquinolones, the sulfonamide group, and aminoglycosides. Specifically, resistance rates of 100% (n=29) were noted for 2nd generation cefuroxime, 3rd generation ceftriaxone, cefoperazone/sulbactam, and amoxicillin-clavulanate from the aminopenicillin group. However, 96.55% (n=28) exhibited resistance to 5^th-generation cephalosporin ceftolozane/tazobactam, followed by 3^rd-generation ceftazidime at 65.51% (n=19), 2^nd-generation cephalosporin cefoxitin, and sulfonamide group trimethoprim/sulfamethoxazole at 55.17% (n=16) each. Additionally, 51.72% (n=15) resistance was observed for 4th generation cephalosporin cefepime, aminoglycosides amikacin, and fluoroquinolones levofloxacin each. For penicillins plus β-lactamase inhibitors, piperacillin-tazobactam and ticarcillin/clavulanate exhibited 37.93% resistance (n=11). Carbapenems showed resistance of 31% (n=9) for imipenem and 27.58% (n=8) for ertapenem. Although it is uncommon for E. coli to exhibit higher resistance to imipenem than to ertapenem, our data, confirmed through double-checking, indicate this pattern in strains 21380, 22684, and 21501. Further investigation into the resistance mechanisms in these strains may provide insights into this unusual observation ( Figure 1 ). Glycylcycline showed a resistance of 20.68% (n=6), while nitrofurantoin exhibited resistance in 10.34% (n=3) of the cases. Notably, the isolates demonstrated 100% sensitivity to fosfomycin and tobramycin ( Figure 1 ). Overall, the results indicate widespread resistance among the clinical isolates, particularly against cephalosporins. Out of the total 29 isolates, 68.96% (n=20) exhibited a Multidrug-Resistant (MDR) phenotype, indicating resistance to multiple classes of antibiotics ( Figure 1 ). ESBL production was verified in all isolates through the DD-Synergy test. Specifically, 100% of the isolates (n=29) exhibited ESBL production ( Figure 1 ). Additionally, among these isolates, further analysis identified 10 isolates as MBL producers, accounting for approximately 34.48% (n=10) of the total isolates tested ( Figure 1 ). This highlights a noteworthy subset of isolates within the studied population that possess the MBL phenotype in conjunction with ESBL production.

MLST analysis based on the E. coli MLST website scheme revealed that all 29 isolates in our study belonged to sequence type (ST) 410. This indicates a clonal relationship among the isolates, suggesting a common evolutionary lineage. ST410 is an important sequence type associated with E. coli infections. The consistent identification of ST410 among our isolates suggests a potential endemicity or dissemination of this particular sequence type in our study population. Our dataset comprises 607 E. coli ST410 isolates, with representation from various regions: with the highest representation observed in Thailand (n=98), followed by the United States (n=82), China (n=78), Australia (n=56), Israel (n=56), France (n=52), Germany (n=30), Niger (n=15), Canada (n=19), Qatar (n=18), Switzerland (n=15), Denmark (n=14), Pakistan (n=9), Singapore (n=9), Saudi Arabia (n=8) and others between 2009-2021. The distribution of isolates from other regions is visualized in Figure 2 . Phylogenetic group analysis revealed that the majority of the isolates belonged to group B2 51.72% (n=15), indicating a higher association with extraintestinal pathogenic E. coli (ExPEC) strains. Group D accounted for 24.14% of the isolates (n=7), followed by Group A at 13.79% (n=4) and Group B1 at 10.34% (n=3). The serotyping analysis identified multiple serotypes among the isolates. The most frequently identified serotype was O1:H7 at 27.59% (n=8), followed by O2:H1 at 20.69% (n=6), O8:H4 at 17.24% (n=5), O16:H5 at 13.79% (n=4), and O25:H4 at 10.34% (n=3). Other serotypes identified by 10.34% included O6:H1 (n=1), O15:H5 (n=1), and O18:H7 (n=1) ( Figure 3 ). These findings highlight the distribution and prevalence of phylogenetic groups and serotypes among the isolates in our study population. We established a comprehensive correlation between serotype, phylogenetic group, and specimen sources, as illustrated in Figure 3 . While we observed a diverse distribution of serotypes within phylogenetic group B2, including O1:H7, O2:H1, O16:H5, and O15:H5. our analysis did not reveal a clear or consistent association between specific serotypes and specimen origins. The D group exhibited a distinct serotype profile, comprising O8:H4, O18:H7, and O6:H5. Remarkably, specimens originating from blood, urine, pus, and stool were associated with this phylogenetic group. Phylogenetic group A demonstrated a specific serotype pattern, with isolates exhibiting O2:H1 and O16:H5, all sourced exclusively from urine specimens. Additionally, isolates in phylogenetic group B1 displayed the O25:H4 serotype, with specimens originating from both blood and urine sources. Despite these observations, no definitive pattern of serotype distribution relative to specimen source was established. These results indicate that while trends in serotype distribution were noted, a significant correlation between specific serotypes and specimen types was not evident. Further investigation may be necessary to better understand the factors influencing serotype distribution and specimen associations.

The most commonly detected ESBL genes among the isolates were bla _CTX-M 89.65% (n=26), indicating the higher presence of extended-spectrum β-lactamase enzymes. The specific types identified were bla _CTX-M-15 34.48% (n=10), bla _CTX-M-14 20.69% (n=6), bla _CTX-M-27 17.24% (n=5), bla _CTX-M-1 13.79% (n=4), and bla _CTX-M-140 3.45% (n=1). In addition to bla _CTX-M, other ESBL genes detected included bla _TEM 79.31% (n=23) and bla _SHV 62.07% (n=18), which are also associated with β-lactam resistance. The detailed breakdown of ESBL types is as follows: 12 strains with TEM-type ESBLs, 8 strains with SHV-type ESBLs, and 6 strains with CTX-M-type ESBLs. Among the 29 strains, TEM and SHV enzymes were detected; however, these are not classified as ESBLs stricto sensu unless they confer resistance to third-generation cephalosporins, which can be confirmed by a positive double-disc synergy test. Specifically, three strains showed a positive double-disc synergy test but did not harbor any CTX-M enzymes. The detailed breakdown of ESBL types is as follows: 12 strains with TEM-type ESBLs, 8 strains with SHV-type ESBLs, and 6 strains with CTX-M-type ESBLs. Furthermore, the following antibiotic resistance genes were identified in the isolates: bla _OXA-1 34.48% (n=10), bla _OXA-48 17.24% (n=5), bla _KPC-2 10.34% (n=3), bla _KPC-3 6.89% (n=2), bla _NDM-113.79% (n=4), bla _NDM-5 3.45% (n=1), bla _GES-1 6.89% (n=2), bla _GES-5 3.45% (n=1), bla _CYM-1 1.34% (n=3). To provide a comprehensive overview, of the percentage of isolates carrying each resistance gene. In addition, other antibiotic resistance genes were identified as aminoglycoside resistance genes aac(6’)-Ib 55.17% (n=16), aac(3)-IIa 10.34% (n=3); fluoroquinolone resistance genes qnrB 20.69% (n=6), and qnrS 13.79% (n=4); chloramphenicol resistance gene catA1 21.14% (n=7), glycylcycline resistance gene tet(X3) 13.79% (n=4) ( Figure 4 ; Supplementary Table 3 ).

Virulence genes were categorized into attachment, biofilm formation, and cellular toxicity/lysis. We analyzed their associations with serotypes and phylogenetic groups, despite the limited sample size. No significant association patterns were observed, suggesting a complex distribution of virulence factors across different serotypes and phylogenetic groups. Among the detected genes, the most prevalent virulence gene was fimH, which encodes the type 1 fimbriae adhesin, found in 100% (n=29) of the isolates, indicating its universal presence among the ST410 strains. The papC gene, which encodes P fimbriae facilitating bacterial adherence to host cells, was detected in 82.76% (n=24) of the isolates. The hlyA gene, associated with hemolysin production, was found in 75.86% (n=22) of the isolates. The cnf1 gene, which was identified in 62.07% (n=18) is associated with cytotoxic necrotizing factor production, which can lead to cellular damage while the eaeA gene was present in 41.38% (n=12) of the isolates. This gene is associated with the formation of attaching and effacing lesions and is commonly found in enteropathogenic E. coli strains. The stx1 and stx2 genes, which encode Shiga toxins, were detected in 27.59% (n=8) and 20.69% (n=6) of the isolates, respectively ( Figure 4 ). These toxins are known to be involved in the development of severe gastrointestinal symptoms. In addition to virulence genes, the presence of genes associated with biofilm formation was investigated. The bap gene, associated with biofilm-associated protein production, was found in 10.34% (n=3) of the isolates. Genes involved in curli fiber production, including csgA, csgD, and csgE, were detected in 17.24% (n=5), 13.79% (n=4), and 10.34% (n=3) of the isolates, respectively ( Figure 4 ). These results highlight the presence of various virulence/toxin genes and biofilm-associated genes in the studied isolates. The detection of these genes suggests the potential pathogenicity of the isolates and their ability to cause severe gastrointestinal symptoms and cellular damage. Furthermore, the presence of biofilm-associated genes indicates the ability of the isolates to form biofilms, which can enhance bacterial survival and contribute to antimicrobial resistance. In addition to characterizing the antibiotic resistance and virulence profiles, we conducted an in-depth analysis of the correlation between resistance genes, virulence genes, and biofilm-associated genes. The correlation heatmap, presented in Figure 4 , provides a visual representation of the associations between various resistance genes, virulence factors, and biofilm-producing genes. The correlation heatmap elucidates potential relationships between these gene categories, offering valuable insights into the co-occurrence patterns. This comprehensive analysis enhances our understanding of the interconnectedness of resistance, virulence, and biofilm formation in the studied isolates.

Diverse plasmids were observed among the ST410 isolates, representing a range of plasmid types. The most commonly identified plasmid IncFIS plasmid, known for its broad host range and association with antibiotic resistance genes, was detected in 93.10% (n=26) of the isolates. The IncX4 plasmid, often associated with the spread of multidrug resistance, was found in 72.41% (n=21) of the isolates. The IncFIA, IncCol, IncI2, and IncFIC plasmids, known to carry various resistance determinants, were identified in 65.51%, 51.72%, 44.82% and 41.37% (n=19), (n=15), (n=13), and (n=12) of the isolates, respectively. These plasmids exhibited varying transmission frequencies, with ranges from 1.3X10^-2 to 2.7X10^-3 ( Figure 5 ). The presence of multiple plasmid types suggests the potential for horizontal gene transfer and the dissemination of resistance and virulence genes among the ST410 isolates.