Genomic Evaluation of Multidrug-Resistant Extended-Spectrum β-Lactamase (ESBL)-Producing Escherichia coli from Irrigation Water and Fresh Produce in South Africa: A Cross-Sectional Analysis

Escherichia coli, both commensal and pathogenic, can colonize plants and persist in various environments. It indicates fecal contamination in water and food and serves as a marker of antimicrobial resistance. In this context, 61 extended-spectrum β-lactamase (ESBL)-producing E. coli from irrigation water and fresh produce from previous studies were characterized using whole genome sequencing (Illumina MiSeq). The Center for Genomic Epidemiology and Galaxy platforms were used to determine antimicrobial resistance genes, virulence genes, plasmid typing, mobile genetic elements, multilocus sequence typing (MLST), and pathogenicity prediction. In total, 19 known MLST groups were detected among the 61 isolates. Phylogroup B1 (ST58) and Phylogroup E (ST9583) were the most common sequence types. The six ST10 (serotype O101:H9) isolates carried the most resistance genes, spanning eight antibiotic classes. Overall, 95.1% of the isolates carried resistance genes from three or more classes. The blaCTX-M-1, blaCTX-M-14, and blaCTX-M-15 ESBL genes were associated with mobile genetic elements, and all of the E. coli isolates showed a >90% predicted probability of being a human pathogen. This study provided novel genomic information on environmental multidrug-resistant ESBL-producing E. coli from fresh produce and irrigation water, highlighting the environment as a reservoir for multidrug-resistant strains and emphasizing the need for ongoing pathogen surveillance within a One Health context.


■ INTRODUCTION
Escherichia coli, a gram-negative bacteria, is one of the most intensively studied microorganisms. 1 As a commensal organism, it is among the first colonizing bacteria in the gastrointestinal tracts of humans and animals naturally occurring in the environment (water, soil, plants). 2,3Additionally, at least 11 pathotypes causing disease in humans and animals have been described and are classified into two categories: intestinal pathogenic (IPEC) and extraintestinal pathogenic (ExPEC) E. coli. 4,5The pathotype differentiation is based on the presence of specific virulence factors, mechanisms of infection, and interactions with host cells. 5Furthermore, E. coli strains belong to different phylogenetic groups, which are intertwined with virulence factors and the genetic substructures associated with different phylogeny, phenotypic, and genotypic traits. 6,7−10 Typically, E. coli infections among humans are associated with phylogroups B2 and, to a lesser extent, D, while phylogroups A and B1 are often associated with commensal E. coli. 8,11he IPEC pathotypes causing disease in humans and animals include enteropathogenic E. coli (EPEC), enterohemorrhagic/Shiga toxin-producing E. coli (EHEC/STEC), enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAEC), diffusely adherent E. coli (DAEC), enteroinvasive E. coli (EIEC), and adherent invasive E. coli (AIEC). 12Generally, foodborne disease outbreaks have been associated with IPEC pathotypes, particularly EHEC/STEC. 13The characteristic virulence factors responsible for associated clinical symptoms of IPEC easily distinguish them from commensal E. coli; however, distinguishing ExPEC is difficult. 14Variants within the ExPEC group are classified according to the host and site of infection as uropathogenic E. coli (UPEC), neonatal meningitis E. coli (NMEC), avian pathogenic E. coli (APEC), and septicemia-associated E. coli (SEPEC). 11,12The virulence factors of ExPEC may be present in various combinations and can be divided into five main categories, namely, ironsequestering systems (iucD, irp2, and chuA), adhesins (papC, F10papA, sfaDE, afaBC III, iha, f imH, clpG, tsh, and hra), invasin (ibe10), protectins (TraT, OmpA, and the capsular antigen K), and toxins (ompT, ehxA, espP, hlyA, hlyD, vat, sat, and cnf 1). 12,13ExPEC infections are being recognized as an emerging serious public health threat due to the increased acquisition of new and troubling antibiotic-resistance genes, leading to ineffective treatment options. 13E. coli is globally reported as one of the leading pathogens responsible for human deaths associated with antimicrobial resistance. 15,16s E. coli, both commensal and pathogenic, can colonize and persist in various niches, it is often used as an indicator of fecal contamination in water and food safety system monitoring. 17 UPMP 2050

SAMN19374582 fresh produce
not tested for phenotypic resistance UPMP 2120

Environmental Science & Technology
More recently, it is also an indicator of antimicrobial resistance dynamics in a One Health context due to its genomic plasticity and frequent exposure to antimicrobial pressure. 18,19Indeed, the One Health paradigm implies that human and animal health and the environment are interdependent. 18Potential reservoirs for ExPEC include nonhuman reservoirs such as surface water, food animals, fresh produce, soil, sewage, and wastewater effluent. 15,20The ubiquity of E. coli renders it a One Health problem involving the water-plant-animal-foodpublic health interface; therefore, standardizing surveillance methodology across all reservoirs becomes important to be able to produce reliable, comparable data of the circulating genomic background. 21hole genome sequencing (WGS) has become the tool of choice in laboratory-based outbreak investigations, particularly in public health. 21,22In addition to public health surveillance, within a food safety context, many high-income countries have successfully adopted WGS in routine food surveillance/ monitoring systems. 21,23Higher accuracy insight into isolate relationships is provided with WGS analysis, making it possible to track trends associated with pathogen virulence and antimicrobial resistance.This can support risk assessment when combined with available metadata across all One Health domains. 21However, Richter et al. 24 recently reported that the use of WGS in environmental surveillance studies in low-and middle-income countries (LMICs) remains low.
It is well documented that potential microbial contamination arises along many points throughout fresh produce production and supply systems. 25In South Africa, the dualistic fresh produce production system consists of highly regulated formal systems with commercial farms as well as the informal system, where predominantly small-scale farmers often have limited resources and infrastructure. 26However, across all fresh produce production in South Africa (both formal and informal), agricultural irrigation water sources predominantly include surface water (rivers, streams, dams, and canals) as well as borehole water. 28−30 Typically, water used for irrigation will either be directly applied to the field from the specific water source or pumped into a holding dam or water reservoir until use. 30,31The current study aimed to evaluate the circulating antimicrobial resistance genes, virulence factors, and serotypes of 61 historically isolated multidrug-resistant ESBL-producing E. coli (2016−2019) from water and fresh produce samples in South Africa, 26−33 using WGS.Furthermore, to establish baseline genomic information on the predicted pathogenicity of environmental isolates comparable to existing clinical data.

■ MATERIALS AND METHODS
Multidrug-Resistant ESBL-Producing E. coli Selected for Whole Genome Sequence Analysis.Sequences of 61 multidrug-resistant ESBL-producing environmental E. coli isolates were retrieved from the National Center for Biotechnology Information (NCBI) GenBank database under the BioProject accession number PRJNA642017 for in-depth genomic characterization (Table 1).The de novo assembly metrics of all sequences are shown in the Supporting Information Table 1.The contigs were subsequently submitted to the Galaxy platform (https://usegalaxy.eu/),Center for Genomic Epidemiology (CGE) platform (https://cge.cbs.dtu.dk/services/), and Technical University of Denmark (DTU) for bioinformatics analysis.
Phylogenetic Screening.All genomes were annotated using Prokka (Galaxy Version 1.14.6 + galaxy1), 34 and the E. coli core genome alignments were constructed using Roary (Galaxy Version 3.13.0+ galaxy2) 35 based on the genome annotation files (gff3 file).The default parameters (95% identity for blastp and 99% of isolates a gene must be in to be core) were used in Roary to classify the core/unique genes.Subsequently, the "core gene alignment" Roary results were used to construct a phylogenetic tree using Fasttree (Galaxy Version 2.1.10+ galaxy1) and visualized using iTOL. 36A core

Environmental Science & Technology
genome MLST (cgMLST) analysis was additionally performed with the default settings using cgMLSTFinder-1.2on the CGE platform 37,38 and visualized in iTOL.The minimum spanning tree from the E. coli isolates based on the MLST scheme was generated using SeqSphere+. 39The different phylogroups of the E. coli isolates were determined using in silico ClermonTyping. 9ene Screening.Using the CGE platform (https://cge.−42 Default parameters were considered for all of the software used unless otherwise indicated.With ABRicate (https://github.com/tseemann/abricate), the AMR gene presence was corroborated using the Comprehensive Antibiotic Resistance Database (CARD), ARG-ANNOT, Res-Finder, NCBI AMRFinder Plus, and MEGA Res databases, 43−48 while the presence of metal resistance genes was determined with BacMet version 2.0. 49Furthermore, mobile genetic elements and their association with virulence and antimicrobial resistance genes were determined with MobileE-lementFinder (Version 1.0.3), 50and the presence of integrons with IntegronFinder version 2.0, 51 while PathogenFinder version 1.1 was used to predict the pathogenicity of the E. coli isolates toward human hosts. 52RESULTS Phylogroups, Sequence Types, and Serotypes of the E. coli Isolates.The phylogenetic grouping showed that E. coli belonging to phylogroups A, B1, C, D, E, F, and G were recovered from water samples, and E. coli that belonged to phylogroups A, B1, B2, D, E, and G were recovered from fresh produce samples.Of the 61 E. coli isolates, phylogroups A (31.15%) and B1 (27.87%) were the most common in the environmental samples.The B2 isolates (3.28%) were recovered from fresh produce samples only, while isolates belonging to phylogroups D (6.56%) and E (14.75%) were recovered from water and fresh produce samples.Interestingly, four isolates (6.56%) from both water and fresh produce samples belonged to phylogroup G, which is closely related to phylogroup B2. 10 A total of 19 known MLST groups were detected among the 61 isolates (Figure 1 and Tables 3−5).ST58 (n = 10, 16.39%), belonging to phylogroup B1, and ST9583 (n = 8, 13.11%), belonging to phylogroup E, were the most common E. coli sequence types associated with the environmental (water and fresh produce) samples.Within phylogroup B1 isolates, other STs found included ST162, ST602, and ST847.The ST10 (n = 6, 11.48%) isolates were restricted to the water samples and detected only in isolates belonging to phylogroup A (Figure 1).Other STs associated with phylogroup A were ST48, ST93, ST226, ST681, ST752, and ST1585.ST1193 was detected in phylogroup B2 and ST117 in phylogroup G. Four isolates (6.56%) within phylogroups B1, E, and F belonged to unknown STs, while most of the other MLST groups were detected to a limited extent (Figure 2).Using Enterobase, 53 the unknown STs were identified within the MLST-Enterobase (ST210d7d18a802c59df81880a978149a02c49a6021b, STc75-778699e0a2b1faca8b5d6f9051eb7d9defca4, and ST85e7b10e-b1371e1fae7d8bf12c0066e6a995add0) and MLST-Pasteur (STb0618816d6163930f5c1952a39b99044904119f5, ST782-9ecb9c01fd0f5134a9452e5ded95cdbc670dc, and STcddffc61-67c10ddd07704ddd888c485cedb717d2) databases.

Environmental Science & Technology
No distinct pattern was observed among the MLST groups, phylogroups, and serotypes (i.e., O-and H-typing).A total of 14 (22.95%)isolates had an untypeable O-type, while the Htypes for these isolates varied between H19, H2, H21, H37, and H39.One isolate from fresh produce belonging to phylogroup A was determined to contain two O-types on the same contig (Figure 2).Eight isolates (ST10, ST753, or ST1585) with the O101 serogroups had either H9 or H10 types and belonged predominantly to phylogroup A (Figure 2).In total, a diversity of 27 serotypes were detected across the complete collection (Figure 2).
Characteristics of Virulence Factors among the E. coli Isolates.The virulence genes detected belonged to the adhesins, nutritional/metabolic, biofilm, invasion, and effector delivery systems virulence factor categories. 54 The distribution of the different virulence genes mostly depended on the phylogroups and sequence types (Figure 2).No distinct difference was seen in virulence-associated genes detected in E. coli isolated from water compared to those isolated from fresh produce samples.The most frequent virulence genes identified within the adhesin category were the UPEC-associated f imH, followed by the APEC-associated ipfA.Within the nutritional/ metabolic category, the iron uptake virulence genes were predominant in isolates that belonged to phylogroup G (Figure 2).Furthermore, the chuA virulence factor was the only gene that regulated iron uptake detected in isolates belonging to phylogroup E. The UPEC-associated traT protectin virulence factor was present in 81.97% (n = 50) of the isolates.Except for one phylogroup A isolate where espA, espB, and espF were found, the toxin category genes (ompT, n = 2 and vat, n = 4) were exclusively present in isolates from phylogroups B2 and G, respectively (Figure 2).
Antimicrobial Resistance Genes.Only 95.1% of the 61 isolates presented a potentially multidrug-resistant genotype with resistant genes in three or more antibiotic classes present (Figure 3).Overall, 60.66% of the E. coli isolates had aminoglycoside resistance genes present.The most common aminoglycoside-modifying enzyme encoding genes were aph-(6)-Id followed by ant(3″)-Ia, aadA2, and aph(3″)-Ib.The phenotypic antimicrobial resistance patterns of the E. coli isolates from the individual cross-sectional studies 26,27,30−33 showed that at least 55 of the E. coli isolates exhibited a phenotypic multidrug resistance profile, with resistance to different antibiotics in at least three different antibiotic classes (Table 2).
Mobile Genetic Elements.Mobile genetic elements associated with similar virulence and antimicrobial resistance genes were present in E. coli isolates from water and fresh produce samples (Table 3).Interestingly, only water sample E. coli isolates had Inc.FII and Inc.FII(pRSB107) plasmids associated with aminoglycoside, tetracycline, and chloramphenicol antimicrobial resistance genes (Table 4), while E. coli from water and fresh produce samples carried the same plasmids (Inc.FII and Inc.FII(pRSB107)) with associated virulence factors.The ESBL-encoding genes associated with mobile genetic elements were bla CTX-M-1 (associated with ISEc9 and Inc.1), bla CTX-M-14 (associated with IS26), bla CTX-M-15 (associated with ISEc9 and ISKpn19), and bla CTX-27 (associated with IS102) (Tables 3−5).In total, 31/61 isolates did not contain integrons, while three types of elements (complete integron, In0, and CALIN) were identified in the remaining isolates.IntegronFinder distinguishes a complete integron as an integron with an integron integrase nearby attC sites.From the current study, three isolates from leafy green vegetables and 12 isolates from river, borehole, or dam water contained complete Class 1 integrons.An In0 element is distinguished as an integron integrase only, without any attC site nearby, and three isolates from cucumber, spinach, and canal water carried In0 elements.CALIN elements are described as clusters of attC sites lacking integrase nearby or a degraded integron.In the current study, nine isolates from dam water, cabbage, spinach, and apple samples carried CALIN elements.Overall, all of the E. coli isolates showed a > 90% predicted probability of being a human pathogen.This follows as the PathogenFinder tool provides a fast estimation of the pathogenic potential of bacteria based on the identification of gene families that correlate with pathogenicity in known and unknown strains. 52

■ DISCUSSION
To the authors' knowledge, this is the first study presenting genomic information on environmental multidrug-resistant ESBL-producing E. coli isolates from fresh produce and irrigation water sources in South Africa.In total, 59% of the multidrug-resistant ESBL-producing E. coli were considered commensal based on the phylogroups.It is well known that E. coli is ubiquitous and forms part of the natural flora of the gastrointestinal system of humans and animals.Apart from the traditional virulence factors and toxins that define pathogenicity, molecular features such as the ability to evade the host's immune system or a group of genes to activate other genes also contribute toward bacterial pathogenesis. 52In the current study, PathogenFinder was used to predict the probability of environmental E. coli being a human pathogen.The pipeline matched the genomic input against known pathogenic and nonpathogenic gene families as the presence of gene families containing proteins with unknown functions has also been reported to play an important role in pathogenicity, 52 resulting in all isolates having a > 90% predicted probability of being a human pathogen.Commensal E. coli with no pathogenic features, as well as intestinal pathogenic strains, are most often observed in phylogroups A or B1. 54Correspondingly, most of the E. coli isolates from the current study belonged to phylogroups A and B1; however, no virulence genes associated with intestinal pathogenic strains were present.
Notably, ten E. coli ST58 strains belonging to phylogroup B1 were detected in the current study.However, the organisms' ability to acquire both resistant determinants and virulence factors results in harmless commensals becoming emerging human pathogens, capable of causing a broad spectrum of intestinal and extraintestinal disease. 55,56Previously, E. coli ST58 harboring multiple antimicrobial resistance and virulence genes have been reported in store-bought fresh produce as well as from pork sausage in Germany, 57,58 similar to the results from the current study.Although limited information is available about the ST58 serotypes detected in the current    59 In foodproducing environments, E. coli is often used to indicate fecal contamination as it appears at low background levels in the environment but has high survival rates. 60Furthermore, the WHO reported that ESBL-producing E. coli should be used as an indicator in monitoring programs to facilitate the establishment of integrated multisectoral antimicrobial resistance surveillance in One Health. 19Interestingly, in four isolates (two phylogroup A ST93 and two phylogroup B1 ST847 and ST58 E. coli isolates) combinations of KpsMII_K5, iutA, and papC virulence factors, among others, were present.According to Johnson et al., 61 for isolates to be classified as ExPEC, two or more of the papAH, and/or papC (P-fimbriae), sfa-focDE (S-and F1C-fimbriae), afa-draBC (Dr-binding adhesins), iutA (aerobactin siderophore system), and kpsMII (group 2 capsules) virulence factors need to be present.Other strains from the current study that also harbored two or more virulence factors for the acknowledged molecular definition of ExPEC belonged to phylogroups D (from water and fresh produce samples) and B2 (from fresh produce samples).Moreover, four strains from water and fresh produce samples from the current study belonged to phylogroup G. Clermont et al. 10 reported that phylogroup G strains are highly virulent with antibiotic-resistance potential and are closely related to phylogroup B2.These strains represent around 1% of E. coli in humans and, although uncommon, have previously been isolated from livestock, poultry, and poultry meat in the East of England and Northern Europe. 10n the current study, all phylogroup G strains belonged to the ST117 lineage, previously reported as the most prevalent lineage in phylogroup G and reported as a poultry-associated lineage with the ability to also establish in humans and cause severe extraintestinal diseases. 10From the current study, the phylogroup G ST117 isolates were obtained from irrigation water and fresh produce samples, and all four strains harbored the ExPEC determining virulence factors.Typically, E. coli strains responsible for extraintestinal infections belong to phylogroup B2, D, and F. 61,62 The phylogroup D isolates from this study all belonged to ST69 lineages.Recently, ExPEC ST69 has been reported among the major lineages globally ("top 20 commonest ExPEC sequence types") 63,64 and has been isolated from raw vegetables in South Korea 65 as well as from poultry and humans in Zambia. 66n contrast to previous studies, the four E. coli ST69 strains from the current study had different serotypes and did not harbor plasmids associated with antimicrobial resistance genes.Other common lineages among ExPEC include ST10, and in the current study, six of the phylogroup A E. coli isolates were characterized as the O101:H-ST10 strains.−68 Typically, serotype O101 is detected among pathogenic E. coli, associated with animal and human disease, with serotype O101:H9 predominantly reported in Shiga toxin-producing E. coli (STEC).Interestingly, the O101:H9-ST10 strains from the current study did not harbor any stx1, stx2, eaeA, or ehxA virulence genes usually associated with STEC; 5 however, antimicrobial resistance genes from at least eight different classes were present among these strains.
Although limited studies have focused on the surveillance of nonpathogenic bacteria, the significance of commensals as reservoirs of antimicrobial resistance in the environment and food chains is gaining more attention. 15,69As an example, Gekenidis et al. 70 reported on the occurrence of antibioticresistant environmental E. coli from drain water and irrigated chive plants through a complete irrigation chain with resistance determinants for up to six different antibiotic classes present.Although no clear distinction was seen between the resistance profiles of E. coli from irrigation water versus those of E. coli from fresh produce in the current study, the phylogroup E and G strains generally harbored fewer resistance genes than isolates that belonged to the other phylogenetic groups.
From the current study, 95.10% of the environmental strains showed a potential for multidrug resistance based on the genomic profile, with multidrug resistance defined as nonsusceptibility to at least one agent in three or more antimicrobial categories. 71This contrasts with results from a study in Uganda, where the commensal E. coli isolates from food animals, characterized using WGS, harbored a limited Environmental Science & Technology number of antimicrobial resistance genes. 15Notably, none of the isolates in the current study harbored the plasmidmediated colistin resistance gene (mcr) or carbapenemase resistance genes (bla NDM , bla KPC , bla VIM , and bla OXA-48 ).−75 However, multidrug-resistant E. coli isolates harboring clinically significant bla CTX-M genetic determinants, among others, have previously been reported in water 76 and fresh produce 70 samples, which correspond to the results from the current study.Currently, the most prevalent ESBL globally reported in clinical isolates, human and animal fecal matter, and the aquatic environment is bla CTX-M-15 . 71,77The predominant βlactamase resistance genes detected in the current study were bla CTX-M-14 (CTX-M Group 9) followed by bla CTX-M-15 (CTX-M Group 1), and in selected isolates, these genes were associated with insertion sequences.Specifically, in two isolates, bla CTX-M-15 was carried on the insertion sequence ISEc9, which corresponds to a previous study where E. coli was isolated from hospital patients in Nigeria. 78Moreover, the cocarriage of the quinolone resistance gene qnrS1 and bla CTX-M-15 in association with insertion sequence ISKpn19 from the current study corresponds to E. coli characterized from dairy farms in Quebec, Canada. 79The Inc.FII plasmid, known globally to contribute toward the spread of clinically relevant antimicrobial resistance genes, 80 was detected in association with certain virulence and antimicrobial resistance genes in isolates from water and fresh produce samples in the current study.Within a One Health context, these results emphasize the significance of monitoring food-producing environments, including water and  81 reported on environmental E. coli from wastewater treatment plants and receiving river water in Kwazulu-Natal (South Africa) that cocarried antimicrobial resistance, heavy metal (mercury and chromate), and disinfectant (quaternary ammonium compounds) genes.In contrast, isolates from the current study did not harbor any heavy metal resistance genes.However, the biocide resistance qacE gene as well as the sul1 antimicrobial resistance gene, which are typically found at the 3′ conserved segment in a class 1 integron, 82 was present in two isolates from the current study where complete integrons were identified.Similarly, E. coli isolates from wastewater treatment plants in South Africa, 81 broiler chickens in the South of Iran, 83 and human, animal, and environmental samples from countries of the Andean Community 84 have been reported to carry complete class 1 integrons.Integrons carrying multiple antibiotic-resistance genes or virulence genes, embedded within mobile genetic elements, significantly contribute toward bacteria across different One Health sectors acquiring traits through the cotransfer of genes, which can increase pathogenicity. 85t is well documented that interconnected reservoirs of antimicrobial-resistant bacteria include animals, humans, and food, which allows rapid gene exchange through horizontal gene transfer within food systems. 86From a food safety perspective, identifying microbial contaminants in the waterplant-food nexus is vital for hazard characterization.In African countries, including South Africa, the evidence of STEC O157:H7 occurrence in the environment and infection among animals and humans, in general, is not conclusive. 870][31][32][33]87 Although antimicrobial-resistant bacteria complicate food safety assurance, 2 building a genomic database of the virulence genes, antimicrobial resistance genes, and potential pathogenicity of environmental isolates, comparable to existing clinical data, is essential for the implementation of risk mitigation strategies. A imitation of the current study is the use of short sequencing reads, preventing complete plasmid assembly and establishing the role of the detected plasmids in gene transfer among environmental bacteria.A recommendation for future research is therefore to combine phenotypic and long-and short-read whole genome sequencing characterization along with gene transfer studies to be able to investigate the role that plasmids play in mediating resistance within foodproducing environments. The results from the current study highlighted the important role that the environment has as a reservoir of multidrug-resistant E. coli and, furthermore, the critical need for continuous potential pathogen surveillance within a One Health context.Future studies should further explore surveillance of the One Health environment.

Figure 1 .
Figure 1.cgMLST-based minimum spanning tree of 61 E. coli isolates recovered from water and fresh produce from formal and informal fresh produce production systems in South Africa.Isolates belonging to the same dominant sequence types (ST) are circled and labeled, and the isolation source is shown in different colors.

Figure 2 .
Figure 2. cgMLST-based phylogenetic tree showing the distribution of virulence genes by the phylogroup, sequence type (ST), serogroup, and isolation source (water, blue and fresh produce, green) of E. coli.The colored circles indicate the presence (filled) or absence (open circle) of the different virulence genes with Adhesins, Iron uptake genes, Protectins, and Toxins typically associated with ExPEC.

Figure 3 .
Figure 3. cgMLST-based phylogenetic tree showing the distribution of antimicrobial resistance genes by the phylogroup, sequence type (ST), serogroup, and isolation source (water: blue and fresh produce: green) of E. coli.The colored circles indicate the presence (filled) or absence (open circle) of the different genes within the different antibiotic classes.

Table 1 .
Summary of 61 Environmental E. coli Strains Previously Reported in South African Point-Prevalence Studies for which the Short-Read Sequences were Retrieved from the National Center for Biotechnology Information (NCBI) Database for Metadata Whole Genome Sequence Analysis

Table 2 .
Antimicrobial Resistance Results of E. coli Isolates from Water and Fresh Produce Samples in Formal and Informal Production Systems in South Africa

Table 3 .
Mobile Genetic Elements Associated with Virulence and Antimicrobial Resistance Genes in E. coli (Grouped According to Sequence Type) Isolated from Water and Fresh Produce Samples in South Africa

Table 3
TEM-1B ; aph(6)-Id; tet(A); aph(3″)-Ib; sul2 Environmental Science & Technology study, it is well documented that ST58 E. coli strains have caused human extraintestinal infections, including sepsis, and are reported as one of the main ESBL-producing E. coli circulating in the human−animal−environment.

Table 4 .
Mobile Genetic Elements Associated with Virulence and Antimicrobial Resistance Genes in E. coli (Grouped According to the Sequence Type) Isolated from Water Samples only in South Africa

Table 5 .
Mobile Genetic Elements Associated with Virulence and Antimicrobial Resistance Genes in E. coli (Grouped According to the Sequence Type) Isolated from Fresh Produce Samples only in South Africa

■ ASSOCIATED CONTENT * sı Supporting Information The
Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.4c02431.The assembly metrics of the whole genome sequences of E. coli isolated from water and fresh produce samples in South Africa (XLSX) Department of Plant and Soil Sciences, University of Pretoria, Pretoria 0001, South Africa; Department of Science and Innovation, National Research Foundation Centre of Excellence in Food Security, Bellville 7535, South Africa; orcid.org/0000-0003-0232-7659;Email: lise.korsten@up.ac.za