Characterizing the antibiotic resistance and virulence genes of diarrheagenic Escherichia coli from Brazilian beef cattle

Abstract

Diarrheagenic E. coli (DEC) can cause severe diarrhea, hemorrhagic colitis, and hemolytic uremic syndrome (HUS). Because DEC is frequently found in ruminants, we investigated the genomic characteristics of different E. coli pathotypes collected from Brazilian beef abattoirs. Pathotypes included Shiga toxin-producing E. coli (STEC), enterohemorrhagic E. coli (EHEC) and enteropathogenic E. coli (EPEC). Bacterial isolates were collected from beef cattle carcasses (n = 3) and bovine feces (n = 28) in 2019. Antimicrobial resistance was documented by agar dilution method for six antibiotics (amoxicillin, ceftiofur, ciprofloxacin, chloramphenicol, sulfamethoxazole + trimethoprim and tetracycline) and later confirmed with genotypic data following whole genome sequencing. The majority of the isolates 29/31 (93%) were antibiotic susceptible although one multidrug-resistant isolate harbored several virulence factors. Among STEC and EHEC isolates (n = 24), stx2 was frequently encountered (21/24). All EPEC isolates were atypical (bfp negative), albeit with clinically important serotypes (O157:H7 and O26:H11). Several enterotoxigenic E. coli harbored genetic markers consistent with STEC and ETEC (n = 7), or EPEC and ETEC (n = 1). Fifteen serotypes were identified including O157:H7, O113:H21, and O26:H11, which have been previously associated with human infections. Phylogenetic analysis showed high similarity among human isolates from other countries, consistent with widespread dissemination. The integration of whole genome sequencing and epidemiological surveillance helped to identify the major pathotypes from Brazilian beef abattoirs, STEC and EHEC, naturally present in food producing animals harboring innumerous virulence factors.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-05239-6.

Introduction

Diarrheagenic E. coli (DEC) are grouped into six major categories: enteropathogenic E. coli (EPEC), enteroinvasive E. coli (EIEC), enteroaggregative E. coli (EAEC), diffusely adherent E. coli (DAEC), enterotoxigenic E. coli (ETEC), and Shiga toxin-producing E. coli (STEC), the latter including E. coli EHEC [20]. These pathotypes harbor characteristic molecular markers and putative virulence genes, although considerable genetic variation amongst E. coli limits gene-for-phenotype associations [35] and presents combinations of genes from different pathotypes that have been termed hybrids and virulent hybrids [5, 12].

Within DEC pathotypes, STEC is considered a particularly important threat to public health and has been one of the main causative agents of foodborne outbreaks worldwide, with some infections presenting as hemorrhagic colitis and hemolytic uremic syndrome (HUS) [25]. Many food- and water-borne outbreaks and sporadic cases causing human illness have been traced worldwide to consumption of beef products, with E. coli O157:H7 being the most prevalent serotype linked to human cases. Nevertheless, other STEC serotypes are often involved in outbreaks, and more than 100 serotypes associated with disease in people [27, 70].

Given that Brazil is one of the world’s largest beef producers, and there is insufficient information about the characteristics of STEC strains circulating in animals in Latin America, especially non-O157:H7 strains [31]. The comparative genomic analysis of Enterohemorrhagic Escherichia coli (EHEC) strains isolated from clinical outbreaks and animal reservoirs in the Americas is critical. Currently, there is a disproportionate scarcity of genomic data from South American lineages [14]. In Brazil, the prevalence and distribution of pathogenic E. coli in cattle has been examined mostly in the state of São Paulo with limited assessments in other states [23]. Expanding this database is essential to elucidate the microevolutionary trajectories, virulence profiles, and antimicrobial resistance patterns that distinguish regional strains. This study was designed to characterize the genomic architecture of DEC strains from cattle sampled previously [24].

Materials and methods

Diarrheagenic E. coli strains

Thirty-one diarrheagenic E. coli isolates, previously identified from bovine slaughterhouses in Brazil [24] were included in this study. The isolates represented three E. coli pathotypes (STEC = 19, EHEC = 9, EPEC = 3) and were selected for whole genome sequence based on their preliminary classification as DEC strains using PCR assays conducted in a previous study, collected from two slaughterhouses, inspected by the Ministry of Agriculture, Livestock and Food Supply (MAPA, Brasília, DF, Brazil), located in two Brazilian states (São Paulo and Mato Grosso do Sul) [24]. The isolates were maintained at − 80 °C in brain heart infusion broth (BHI, Oxoid, Basingstoke, England) supplemented with 20% (v/v) glycerol (Merck, Whitehouse Station, NJ, USA).

Phenotypic resistance

Agar dilution assays were used to assess antibiotic susceptibility against amoxicillin (8 ug/ml), ceftiofur (8 ug/ml), ciprofloxacin (1 ug/ml), chloramphenicol (32 ug/ml), sulfamethoxazole + trimethoprim (4/76 ug/ml) and tetracycline (16 ug/ml). Escherichia coli ATCC 25922 was used as a susceptible control. Assays were conducted as described by CLSI [18] with modifications described by Andretta et al. [2], and breakpoint concentrations were from CLSI [17], and EUCAST [26] following the specific guidelines for the chosen antibiotics. Briefly, isolates were inoculated into 5-ml BHI and incubated overnight (35 °C ± 2 °C). Cultures were adjusted to an optical density equivalent to a 0.5 McFarland standard before being arrayed in a 96-well plate (200 µl per well). Mueller–Hinton agar plates were prepared and supplemented with one of the six antibiotics at their specific breakpoint concentrations (a no-antibiotic plate was used to confirm viability). Broth (2 µL) from the 96-well plate was transferred to the agar plates using a 96-pin replicator (Boekel Scientific, Feasterville, PA, USA) followed by 20 h incubation (35 °C ± 2 °C) in a stationary incubator. Strains that produced a colony in the presence of an antibiotic were considered resistant.

Genome sequencing

A pure culture of each isolate (n = 31) was obtained by streaking the isolates onto MacConkey agar (Oxoid) prior to the sequencing using a NextSeq platform (Illumina, San Diego, CA, USA). DNA extraction and whole genome sequencing (WGS) was carried out by MicrobesNG (Birminghan, England) and Neoprospecta Microbiome Technologies (Florianópolis, SC, Brazil). Reads were processed with Trimmomatic [11], and the clean reads were de novo assembled using SPAdes software [6]. Assembled contigs were annotated using Rapid Prokaryotic Genome Annotation (PROKKA) v.1.13.3 [61]. Kraken 2 was used to check taxonomy [73] and NCBI Foreign Contamination Screen (FCS) was used to remove contaminant sequences in genome assemblies [4].

In silico analysis of WGS data was conducted in different softwires: MLST software (version 2.0.9) was used to identify multi-locus sequence types (MLST) [40], while putative serotypes were identified using the SerotypeFinder [36], and putative virulence genes were identified using VirulenceFinder and virulence factor database VFDB [45, 46]. The VF databases consist of different virulence factors categories including LEE-encoded TTSS effectors, non-LEE encoded TTSS effectors, regulation, secretion system, toxin, adherence, autotransporter, invasion and iron uptake. ABRicate was also used with NCBI AMRFinderPlus database to identify antibiotic resistance genes [28, 60]. Phylogrouping was accomplished using the Clermon Typing Tool [7], and putative mobile genetic elements were identified using the mobile element finder v1.0.3 tool, with minimum alignment coverage of 95%, and minimum sequence identity of 90% [37].

Phylogenetic analyses

The annotated genomes of the Brazilian DEC isolates from bovine feces (n = 3) and beef cattle carcasses (n = 28) and available DEC genomes from previous studies from humans and cattle (n = 85, including the reference sequence E. coli Sakai O157:H7, accession number: SAMN01911278) being the most prevalent serotype linked to human cases, was used to construct a phylogenetic tree based on a distance matrix generated from single nucleotide polymorphisms (SNPs) of core genome genes (Roary software) [54]. The final alignment was processed in ModelTest-NG to select the best substitution model (GTR + I + G4) for the tree reconstruction [22]. RAxML was used to estimate phylogeny [65] and resulting phylograms were further annotated using iTol v 6.8.2 [44]. Comparative E. coli genomes (n = 85) were downloaded from Enterobase [74] and National Center for Biotechnology Information (NCBI) [59]. These isolates were filtered to confirm they were pathotypes of interest (STEC, EHEC, EPEC and hybrid STEC) from cattle or people, being their sequences deposited in NCBI (PRJNA962899 and PRJNA1289705) (Supplementary Table 1).

Results

Phylotype, pathotype and serotype

A total of 31 diarrheagenic E. coli strains were sequenced. This collection represented different E. coli pathotypes (STEC = 19, EHEC = 9 and EPEC = 3), and were distributed among phylogroups E (10/31), B1 (10/31), clade I (6/31), D (2/31), A (2/31), and F (1/31). The average estimated genome size was 5,285,913 bp (range 4,973,916 to 5,556,583), with B1 isolates representing the smaller genomes (average 5,129,301 bp), while phylogroups E and Clade I represented the largest (average 5,393,203 and 5,495,615 bp, respectively).

In silico analysis identified six STEC isolates (stx positive) and one EPEC (eae positive) isolate that harbored sta1, which encodes a heat-stable enterotoxin characteristic of the ETEC pathotype; this mixed combination of pathovar markers is consistent with what has been termed as hybrid isolates [42]. Further, from the nine EHEC O157:H7 isolates, three were characterized as EPEC because they lacked stx by in silico analysis (Table 1).

Table 1.

Brazilian DEC isolates sequenced in the present study

ID	Serogroup	Pathotype PCR	Pathotype WGS	Phylogroup	Genome size
DEC_01	O178:H19	STEC	STEC	B1	4,973,916
DEC_10	O175:H7	STEC	STEC	E	5,157,481
DEC_119	O130:H11	STEC	STEC	B1	5,060,620
DEC_124	O157:H7	EHEC	EHEC	E	5,426,493
DEC_127	O157:H7	EHEC	EHEC	E	5,414,223
DEC_132	O77:H45	STEC	STEC	A	5,122,976
DEC_139	O113:H21	STEC	STEC	B1	5,190,677
DEC_183	O93:H46	STEC	STEC	B1	5,082,195
DEC_200	O105:H28	STEC	STEC	F	5,011,152
DEC_218	O77:H45	STEC	STEC	A	5,176,233
DEC_255	O26:H11	EPEC	EPEC	B1	5,267,551
DEC_258	O26:H11	EPEC	EPEC	B1	5,284,534
DEC_269*	O153:H19	EPEC	ETEC/EPEC	B1	5,066,096
DEC_276	O157:H7	EHEC	EHEC	E	5,457,461
DEC_41*	O2:H45	STEC	ETEC/STEC	Clade I	5,556,583
DEC_48*	O2:H25	STEC	ETEC/STEC	Clade I	5,481,598
DEC_54*	O2:H25	STEC	ETEC/STEC	Clade I	5,517,203
DEC_55*	O2:H25	STEC	ETEC/STEC	Clade I	5,502,886
DEC_56	O2:H15	STEC	STEC	D	5,048,482
DEC_58	O157:H7	EHEC	EHEC	E	5,416,385
DEC_61	O7:H10	STEC	STEC	D	5,305,740
DEC_63*	O2:H25	STEC	ETEC/STEC	Clade I	5,458,400
DEC_64*	O2:H25	STEC	ETEC/STEC	Clade I	5,457,022
DEC_66	O79:H28	STEC	STEC	B1	5,103,868
DEC_71	O113:H21	STEC	STEC	B1	5,115,740
DEC_73	O113:H21	STEC	STEC	B1	5,147,816
DEC_89	O157:H7	EHEC	EHEC	E	5,445,703
DEC_91*	O157:H7	EHEC	EPEC	E	5,378,219
DEC_92	O157:H7	EHEC	EHEC	E	5,446,419
DEC_95*	O157:H7	EHEC	EPEC	E	5,377,223
DEC_96*	O157:H7	EHEC	EPEC	E	5,412,427

^* isolates characterized as different pathotypes based on PCR and WGS

Among STEC isolates, ten serotypes were identified including O178:H19, O175:H7, O130:H11, O113:H21, O93:H46, O105:H28, O77:H45, O2:H15, O7:H10, and O79:H28. EPEC isolates were represented by O26:H11 and O157:H7. All EHEC isolates were represented by the O157:H7 serotype and phylogroup E. Hybrid isolates were divided in three different serotypes with ETEC/STEC being O2:H25 and O2:H45, and ETEC/EPEC being O153:H19.

Phenotypic and genotypic antimicrobial resistance

All DEC isolates (n = 31) were previously tested against six antibiotics that represent antibiotic classes commonly used in farm animals and people in Brazil. Only two of the 31 isolates, STEC DEC 61 and STEC DEC 119, presented phenotypic resistance to different antibiotics, the first presenting resistance to amoxicillin, chloramphenicol, tetracycline, and sulfamethoxazole + trimethoprim, and the second presenting resistance to only amoxicillin. In silico screening confirmed that STEC DEC 61 harbored genes that confer resistance to beta-lactams (bla_TEM-1B), sulfamethoxazole (sul2), trimethoprim (dfrA8), doxycycline and tetracycline [tet(A)], cephalosporin (bla_EC-8), chloramphenicol (catA1) and streptomycin resistance [aph(3'')-Ib and aph(6)-Id]. Further analysis indicated that although some of the genes were found on the same contig (aph(3'')-Ib, aph(6)-Id and sul2- contig 131), none appeared to be associated with plasmids or transposons.

Remaining isolates (n = 29) were considered antibiotic sensitive by phenotypic test, but were positive by in silico analysis for bla_EC alleles (bla_EC, bla_EC-13, bla_EC-15, bla_EC-18 and bla_EC-8). The ETEC/EPEC isolate DEC 269 harbored genes conferring resistance to fosfomycin (fosB), and β-lactams (bla_EC-13). Three isolates (DEC 56, DEC 132 and DEC 269) harbored resistance genes to three or more different antimicrobial classes. Even with the detection of resistance genes, the majority of isolates did not present the respective resistance phenotype using our breakpoint assays.

Virulence gene analysis

Among STEC isolates (n = 19), eight were positive for stx2, three were positive for stx1, and eight harbored both stx1 and stx2. All EHEC isolates (n = 6) harbored only stx2. All EPEC (n = 6) isolates were negative for stx1 and stx2 by WGS analysis, although three of these isolates were previously characterized as EHEC (stx positive) by PCR analysis [24]. All E. coli pathotypes (n = 31) harbored multiple putative virulence genes (Fig. 1). The most common genes of interest included ones that encoded an outer membrane protease (ompT) in (28/31) or 90.3% of the genomes, an outer membrane protein associated with serum resistance in ExPEC (traT) in (27/31) or 87.1% of the genomes, a protein associated with increased serum survival (iss) in (26/31) representing 83.9%, and an adherence protein (iha) in (21/31) representing 67.7%, which is often associated with infective isolates from patients suffering from urinary tract infections [64].

Fig. 1.

Virulence profile of the 31 DEC isolates sequenced in this study. The tree was reconstructed from the core genome genes (n = 13.911) using Roary. The strains are identified by their ID, IDs marked in red are O157:H7 isolates, followed by the phylogroup colors, pathotype (STEC, ETEC/STEC, EPEC/STEC and EHEC) and sequence type (ST). Virulence patterns of genes and phenotypical antimicrobial resistance against different antibiotics: amoxicillin (AMO), ceftiofur (CEF), ciprofloxacin (CIP), chloramphenicol (CLO), sulfamethoxazole + trimethoprim (SUT) and tetracycline (TET), presence (1) or absence (0) are colored in black and white respectively

Most isolates, 77.4% (24/31) harbored a plasmid-encoded enterohemolysin (ehxA). Genes encoding α-hemolysin (hlyCABD) and astA that encodes the enteroaggregative Escherichia coli heat-stable enterotoxin (EAST1), were also present in several isolates (20/31 and 18/31) or 64.5% and 58.1%, respectively. Besides sharing several virulence genes with the EHEC, the STEC isolates that were LEE-negative harbored a unique set of virulence genes (Fig. 1). Multidrug resistant STEC DEC harbored a unique repertoire of virulence genes not found in any other isolates. All EPEC isolates were EAF negative and thus were characterized as aEPEC.

Phylogenetic analysis

The genomes of 116 isolates, comprising 31 DEC isolates from this study and 85 DEC from different pathotypes and geographic regions, were assessed for genetic distance based on SNPs from the core genome. The phylogenetic tree (Fig. 2) shows the isolates grouped according to their respective phylogroups. The O157:H7 isolates from our study clustered together to EHEC O157:H7 reference strain Sakai, one EHEC O157:H7 human strain from Brazil (COPRO21317) and one EHEC O157:H7 bovine strain from Argentina (146N4) on the phylogenetic tree (Fig. 2). The hybrid STEC/ETEC Clade I isolates from our study showed a higher phylogenetic correlation with STEC strains from phylogroup A from our study, while the isolate DEC 269 ETEC/ EPEC strain from phylogroup B1 was closer to STEC and EPEC from Argentina and United States, respectfully. The Phylogroup B1 was very diverse in terms of country of origin and represented majority of the isolates from other countries.

Fig. 2.

Phylogenetic analysis of DEC isolates from Brazil and DEC from other geographic origins. DEC isolates from Brazil (n = 31) are marked with the circles outside the tree and the reference strain E. coli – Sakai O157:H7 is marked with the green star. The inner circle represents the Phylogroups that are indicated by the letters: B1, A, F, E, Clade I, C, and D. The outer circle represents the source of the isolates (Livestock and Human). The tips of the tree are colored according to the country of origin and are color coded following the legend

Discussion

The E. coli genome can vary greatly among different strains, especially in pathogenic E. coli due to genomic evolution and adaptation to different hosts [43]. E. coli can be classified in eight broad phylogroups (A, B1, B2, C, D, E, F) that belong to E. coli sensu stricto and clade I [16]. E. coli clade I shares genetic material and virulence factors with classic E. coli strains but are genetically distinguished from E. coli clades II-V and other isolates from the genus Escherichia [71]. In our study, B1 isolates represented the smaller genomes, while phylogroups E and Clade I represented the largest, also the distribution of phylogroups among the different pathotypes in our study showed that STEC isolates were mostly assigned to phylogroup B1 and all hybrid isolates ETEC/STEC and ETEC/EPEC were assigned to clade I. Although, the higher prevalence of phylogroup B1 in Brazil STEC isolates may be representative of an isolation bias.

The hybrid isolates in our study were all assigned to clade I. Hybrid E. coli can acquire genetic characteristics from other strains [42]. This can happen when molecular markers for ETEC that are often encoded on plasmids are horizontally transmitted to STEC E. coli (genetic markers stx, sta, and/or stb), producing STEC/ETEC hybrids [63]. ETEC are frequently found in cattle, although the prevalence in food can vary from being the most common found pathotype (75%) [66], to relatively rare (2%) [72]. They are generally detected only in clinical isolates, while STEC strains are detected in clinical isolates, as well as in food products [57]. The identification of these hybrid pathotypes should be taken in consideration for epidemiological surveillance [52]. Recently, the emergence of hybrid STEC-ETEC clones have been reported as the cause of diarrhea in humans [5].

The hybrid isolates from our study were categorized in three different serotypes (O2:H25, O2:H45 and O153:H19), consistent with previous studies reporting STEC/ETEC hybrid strains among cattle and people as serogroup O2. Other hybrid strains from clinical cases included other serotypes (O15:H16, O101:H-, O128:H8 and O141:H8) [47, 52]. Given that some hybrid strains have been linked to important outbreaks worldwide, as the cause of Haemolytic Uraemic Syndrome (HUS) and/or potentially fatal systemic infections [8, 58]. Furthermore, it is plausible that the increasing detection of such hybrids in recent studies does not solely reflect an emerging biological phenomenon, but rather the enhanced diagnostic sensitivity provided by advanced genomic technologies and refined surveillance protocols.

EHEC O157:H7 is a worldwide threat to public health and the predominant cause of hemolytic uremic syndrome [56]. Shiga toxins (Stxs) are the main virulence factor of STEC and EHEC isolates, although some EHEC O157:H7 isolates from our study (n = 3) appeared to be missing canonical stx genes based on sequence analysis. The genes encoding Stx are present in chromosomally-integrated prophage that can be lost during in vitro culture. Others have reported the absence of these genes [62]. For example, McCarthy et al., [48] reported that a subset of PCR positive STEC isolates (n = 21) lacked the stx gene during WGS investigation. Ferdous et al., [29] also reported stx negative EPEC strains despite close genetic relationships with STEC. Another possible for the observed loss of the stx gene in the genome could be attributed to the limitations of short-read sequencing technology. While short-read sequencing is cost-effective and highly accurate, an inherent drawback lies in the fragmentation of DNA into small segments, which complicates the assembly of repetitive or complex genomic regions. The stx gene, if located in such a region, might have been missed or misassembled due to ambiguous mapping or insufficient read overlap [1]. In this context, our EPEC isolates O157:H7, were genomically related to EHEC lineages, and may have lost their Shiga toxin genes either during culture in the laboratory or during sequencing.

The stx gene prevalence is variable and depends on multiple factors, while some studies report a higher correlation between STEC and stx2 in isolates obtained from animals [3, 67]. Other studies report high prevalences of stx1 carrying STEC in cattle [41, 53]. The majority of isolates in this study (n = 19) carried stx2 variants. Although, sample size were limited in our study and can influence the final results.

There is insufficient information about the characteristics of STEC strains circulating in animals in Latin America, especially non-O157:H7 strains [31]. Further, non-O157 STEC are emerging as human enteric pathogens, and are often more associated as a group of sporadic cases of illnesses worldwide, with certain serotypes of STEC isolated from beef cattle such as O113:H21 also being implicated with outbreaks causing HUS [34]. Molecular markers that are present in some PAIs of E. coli have been suggested for the identification of LEE-negative isolates (saa, hes, elbG and sab) because these genes are only present in LEE negative STEC isolates [49]. None of these markers were found among our STEC LEE isolates from cattle.

Additionally, some of these genes found in our isolates have been reported in important sequence types (STs) associated with human disease: kpsMII_K5 (group 2 capsule synthesis) is associated with ST131 isolates [21], and is more prevalent in E. coli ST131 from human patients (Mora et al., 2010). Although no ST131 was identified in our isolates, the kpsMII_K5 gene was present in MDR isolate DEC 61 from ST362 and DEC 56 from ST349, both isolates from phylogroup D.

An MDR DEC 61 isolate from this study harbored potentially important genes including afaC, afaD, afaE8, cdtB and hlyF. The afa-3 operon (afaA, afaB, afaC, afaD and afaE8) encodes afimbrial adhesins that are expressed by uropathogenic isolates associated with infections in people and DEC strains are carried by cattle [38]. The presence of the afa operon in pathogenic E. coli is often associated with chronic or recurrent extra-intestinal infections, especially urinary tract infections [69]. cdtB encodes the cytolethal distending toxin B (CdtB), an enzymatically active subunit of the CDT holotoxin [39] among LEE-negative STEC. The presence of cdt has been linked with disease and higher potential to cause HUS [9]. The hlyF gene encodes the hemolysin F and is usually associated with extraintestinal pathogenic E. coli (ExPEC), especially virulent strains of avian pathogenic E. coli (APEC) and neonatal meningitis–associated E. coli (NMEC) [50]. The detection of both extraintestinal and intestinal virulence genes among STEC pathotypes is concerning and the number of hybrid combinations may be increasing, potentially making these pathogens more versatile [30].

kpsE was detected in four (4/31) isolates. This gene encodes a capsule polysaccharide exporting inner membrane protein and is also associated with ExPEC [68]. Similarly, Furlan et al., [30] described virulence genes of diarrheagenic and extraintestinal pathogenic E. coli (ExPEC) in pathogenic E. coli strains from dairy cattle from Brazil, and the presence of hybrid E. coli also harboring antibiotic resistance.

Non-LEE-encoded type III effector genes (nle), such as nleA, nleB, nleC, nleE and nleF were also detected. nleF gene is secreted to the host cells where it reportedly prevents apoptosis in vitro [10]. In the study conducted with EPEC isolates from Brazil, nleF was more prevalent among aEPEC from diarrheal disease patients in Brazil [33]. The nleF gene was only present in one EPEC isolate and one EHEC isolate from our study, although the other genes nleA, nleB, nleC and nleE were present in several of our isolates. These genes, along with nleG2, nleG5, nleG6, nleH1-2 and ent/espL2 have been reported in previous studies to have a higher association with EHEC strains causing HUS and outbreaks in human [19]. Also, Bugarel et al., [13], indicated the gene nleB as possible marker to identify virulent EHEC and EPEC in diagnostic routine.

Although some studies have evaluated the presence of virulence genes, adhesins and toxins in LEE-negative STEC strains isolated from cattle [12, 32], information regarding pathogenic E. coli from farm animals, especially in Brazil, is still scarce [15], and knowledge on the virulence profiles of EHEC O157:H7 and non-O157 helps assess their association with disease. Furthermore, the genome comparison allowed the identification of virulence genes located outside the LEE pathogenicity island that are shared in LEE-negative STEC genomes, but not in the LEE-positive STEC and EPEC isolates, highlighting the potential importance of different virulence genes among pathogenic E. coli. Our isolates harbored a high prevalence (87.0%) of traT and iss, genes associated with STEC belonging to the serogroup O91:H14 from clinical isolates [51] and from avian E. coli isolates from clinical cases and from healthy animals [55].

Conclusions

The integration of WGS and epidemiological surveillance contributes to the monitoring of highly pathogenic E. coli naturally present in food producing animals, facilitating risk assessments and management, with more discriminatory power than conventional typing methods. The identification of hybrid isolates through WGS among DEC isolates from Brazil highlights this important tool in epidemiology studies. These findings can be correlated with virulence gene screening and functional tests to understand the possible link between gene diversification among different serotypes from the same pathotype and the likelihood that some types are more virulent than others. Also, the presence of E. coli from important serotypes (O157:H7, O26:H11 and O113:H21) and extensive virulence genes in isolates from healthy food animals raises the concern of the pathogenic potential of these strains. The overall resistance frequency was low; however, one multidrug-resistant isolate was found to harbor critical virulence factors. Moreover, identification of virulence factors in different serotypes of E. coli from animal origin will be extremely useful in identifying the role of these genes and the contribution of the various factors to the pathogenesis of human disease.

Authors’ contributions

Rafaela de Melo Tavares: Writing – original draft, Software, Investigation, Formal analysis, Methodology, Conceptualization. Ricardo Seiti Yamatogi: Writing – review & editing, Methodology, Conceptualization. Douglas Ruben Call: Writing – review & editing, Formal analysis, Supervision, Conceptualization. Luís Augusto Nero: Writing – review & editing, Resources, Project administration, Investigation, Funding acquisition, Supervision, Conceptualization.

Funding

This work was supported in by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brasília, DF, Brazil), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Brasília, DF, Brazil—Financial code 001) and Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG, Belo Horizonte, MG, Brazil).

Data availability

The datasets generated and/or analysed during the current study are available in the National Center for Biotechnology Information (NCBI), with the BioProject accession numbers PRJNA962899 and PRJNA1289705, and accession codes are indicated in the Supplementary Table 1.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.