Genomic Characterisation of Antibiotic-Resistant Escherichia coli from an Intensive Poultry Production System in the uMgungundlovu District, Kwa Zulu-Natal, South Africa: A Snapshot

1. Introduction

The use of antibiotics in the intensive poultry production industry significantly enhances chicken growth rates, reduces disease prevalence, and provides benefits to both breeders and their flocks [1]. However, administering low doses of antibiotics to chickens for growth promotion creates selective pressure, which leads to the development of antibiotic-resistant strains of commensal bacteria such as E. coli [2]. Antibiotic-resistant bacterial strains potentially spread to humans through occupational exposure or via the food chain [3,4]. Additionally, the transfer of antibiotic-resistant bacteria to soil and aquatic environments has been demonstrated through runoff of chicken litter used as fertilizers [5].

The South African poultry industry is a vital contributor to the country’s agricultural revenue. Poultry meat, the primary source of protein in South Africa, saw consumption reach 2.241 million tons in 2023. This accounted for 18% of the total agricultural gross value [6]. Approximately 23 million chickens are consumed weekly, though this number may vary depending on economic conditions. The rise of antibiotic-resistant E. coli in poultry presents a significant threat to both the economy and the nutritional health of low-and middle-income countries (LMIC), including South Africa [2].

Antibiotics are commonly used in poultry production for growth promotion, disease prophylaxis, and metaphylaxis, and are typically administered orally at varying doses depending on their purpose [3]. From 2002-2004, two-thirds of the 1,500 tons of antibiotics sold for animal use in South Africa were intended for growth promotion in food animals [7]. Recent data from the Department of Health show a marked increase in antibiotic importation for animal use, reaching 2,488,754 tons and nearly doubling the amount used for human disease [8]. The spread of antibiotic-resistant E. coli is facilitated by such use, with potential dissemination through both animal products and environmental runoff. To assess the prevalence and dissemination of antibiotic-resistant E. coli in an intensive poultry production system, E. coli genomes from chicken litter, wastewater, and floor swabs were characterised using whole genome sequencing and bioinformatics analysis to delineate the resistome, mobilome, and phylogeny.

2. Results

2.1. Prevalence of E. coli and Antibiotic Susceptibility Profiling

Out of 54 samples collected, 150 bacterial isolates were randomly selected from chicken litter and wastewater samples to constitute the final sample size, and 47% (n=70) were confirmed as E. coli through molecular screening. Moreover, E. coli was not isolated from any of the floor swab samples. All confirmed E. coli isolates exhibited resistance to at least one antibiotic. Notably, all isolates were resistant to tetracycline (TET). Additionally, 94% (n=66) showed resistance to ampicillin (AMP), while 76% (n=53) were resistant to trimethoprim-sulfamethoxazole (SXT). The lowest resistance rate was observed against ciprofloxacin (CIP), in only 3% (n=2) of isolates. No resistance was detected to gentamicin (GEN), meropenem (MEM), cefotaxime (CTX), ceftriaxone (CRO), ceftazidime (CAZ), or cefepime (FEP) (Table 1). The resistant E. coli isolates consistently exhibited similar resistance profiles across all sampling cycles, except for Cycle 2 (December), during which two isolates showed resistance to ciprofloxacin (Table 1). Ciprofloxacin also had the highest proportion of intermediate susceptibility among the antibiotics tested, with 23% (n=16) of isolates falling into this category. Multidrug resistance (MDR) was defined as, i.e., microbes are resistant to at least one antibiotic in three or more classes [9]. A total of 74% (n=52) isolates were MDR. The most common antibiogram pattern was AMP-SXT-TET accounted for 71% (n=50). AMP-TET-SXT- CIP and AMP-TET-CIP accounted for 1.4% (n=1) each.

2.2. E. Coli Antibiotic Resistance Genes

The whole genome analysis (Table 2) demonstrated that a higher number of isolates 68%, (n=17), carried fluoroquinolone resistance genes in the genotypic analysis as compared to phenotypic AST results. The most frequently detected fluoroquinolone resistance gene was qnrS1, followed by oqxA and oqxB. One isolate (EC04) carried qnrB19, which is linked to plasmid-mediated quinolone resistance. Out of 17 isolates carrying fluoroquinolone resistance genes, eight isolates represented chicken litter, and the remaining nine represented wastewater. A total of 72% (n=18) isolates carried the disinfectant gene sitABC, and 52% (n=13) of sitABC genes were carried on isolates harboring fluoroquinolone resistance.

Extended-spectrum β-lactamase (ESBL) genes were also identified, with multiple bla_TEM genes detected. Bla_{TEM1B -}was the most common gene, followed by blaTEM-135 and the bla_LAP-1 gene. A total of 40% (n=10) of isolates carried ESBL-resistant genes, with an even distribution of five isolates from chicken litter and five from wastewater. These results correspond with AST results. Other resistance genes identified included those conferring resistance to aminoglycosides (aadA1, aadA2, aadA5, aph(3')-Ia, aph(3'')-Ib, aph(6)-Id), bleomycin (bleo), chloramphenicol (cmlA1), fosfomycin (fosA3, fosA4), sulfamethoxazole (sul2, sul3), tetracycline (tetA, tetB, tetM), and trimethoprim (dfrA5, dfrA12, dfrA14, dfrA15, dfrA17). Isolates were further investigated for mutations in quinolone resistance determinant regions (QRDRs). The QRDRs investigated consist of the DNA gyrase (gyrA and gyrB) as well as the DNA topoisomerase IV (parC and parE) genes. In our study, only gyrA (S83L, D87N, D678E*, A828S*, S83L, P872S*), and parC genes (S80I, D475E*, T718A* K665R*) were identified. A total of eight isolates harboured both mutations in the QRDR, while one isolate harbored the gyrA mutation gene only. Among the identified QRDR mutations, there were known mutations with phenotypic effects known to cause resistance to nalidixic acid and ciprofloxacin. Those marked with an asterisk are novel mutations (Table 2).

2.3. Sequence Types and Plasmid Replicons

MLST analysis (Table 2) revealed a variety of sequence types (STs), including ST155, 24% (n=6), ST48, 16% (n=4), ST1286, 12% (n = 3), ST602, 8% (n=2) and the following singletons ST3346, 4% (n=1) ST6050, 4% (n=1), ST359, 4% (n=1), ST1771, 4% (n=1) ST6706, 4% (n=1) ST21, 4% (n=1). The remaining 16% (n=4) of the isolates had unknown STs. All isolates carried plasmid replicons that included IncF, IncFIB, IncFIC, IncFII, IncI2, p0111, IncI1-I(Alpha), IncX1, IncY, Col(pHAD28), IncFIA(HI1), IncFIB(AP001918), IncFIC(FII), IncHI1A, IncHI1B(R27), IncFIA, IncX4, IncQ1, IncB/O/K/Z, IncI(Gamma), Col440I, (pLF82-PhagePlasmid), and ColpVC. The distribution of plasmid replicons among chicken litter and wastewater isolates was even, with Inc being the most common plasmid replicon identified.

2.4. Mobile Genetic Elements Associated with ARGs in the analysed e. Coli genomes

The isolates were further analyzed for ARGs and MGEs on NCBI, and two isolates (EC03 and EC11) fell off due to large genomic size, resulting in a total of 23 isolates (Table 3). The main findings from the synteny analysis (Table 3) illustrated the distribution of MGEs across genomes. Almost all ARGs were associated with insertion sequences (IS) except for EC10; others were associated with transposons except for EC04, EC05, EC07, EC10, EC19, EC25, and EC26, and almost half of the ARGs were carried on integron class 1. A diversity of ISs was evident within isolates; two types of transposons, Tn3 and TnAs1 were observed, whereas integrons were represented by intl class 1 only.

2.5. Phylogenetic Tree Analysis

The phylogenetic tree provides insight into the genetic relatedness of E. coli isolates from South Africa and various African countries (Figure 1). The E. coli genomes from our study showed close genetic relation with poultry derived E. coli from Cameroon, Ghana, Nigeria, and Tanzania; E. coli isolated from the environmental samples, which included unspecified environment isolates, isolated in South Africa, unspecified water from Ghana and South Africa, irrigation water from South Africa, soil from Ghana, and produce from South Africa. These comparisons are based on genomes reported between 2015 and 2024. No data was found for 2025. The isolates appear to have originated from a common ancestor shared with the comparative strains, but somehow branched out of the family tree to form new nodes. Most of the isolates from our study clustered towards poultry isolates; only a few clustered towards the environmental isolates.

2.6. Discussion

The study investigated the phenotypic and genomic characteristics of antibiotic-resistant E. coli from poultry litter and wastewater from an intensive poultry production system. No presumptive E. coli was isolated on all floor swabs sampled after disinfection, indicating an effective sanitation method and less risk of transfer of antibiotic-resistant E. coli strains from one poultry flock to the next production cycle.

2.6.1. Antibiotic Susceptibility Profile

To better understand resistance trends in poultry-associated E. coli, we examined the patterns of antibiotic resistance observed across multiple sampling cycles. The resistance patterns remained largely consistent across the cycles, with all E. coli isolates exhibiting relatively high levels of resistance to three antibiotics, i.e., tetracycline (highest prevalence), followed by ampicillin and trimethoprim-sulfamethoxazole. In the December samples, only two isolates showed resistance to ciprofloxacin EC02 and EC03 (Table 1). The findings align with a study by Phiri et al. [10] in Zambia. An analysis of broiler litter, cloacal swabs, and carcass swabs collected from poultry farms, abattoirs, and open markets across seven districts in Zambia was conducted to assess the prevalence and antibiotic resistance profiles of Salmonella and E. coli. The E. coli isolates showed a high level of resistance to ampicillin (68%), tetracycline (81%), trimethoprim/sulfamethoxazole (65%), and ciprofloxacin, being the least (21%). These results are in contrast to the results found by McIver et al. [11] in a study conducted in a similar region on a commercial poultry farm in South Africa. The study assessed the antibiotic-resistant E. coli obtained from litter, feces, wastewater from the chicken house, truck, crates used in transportation, as well as final meat products for consumption. The AST results demonstrated a low level of antibiotic resistance to AMP, TET, and STX. In comparison, in this study, the low level of fluoroquinolone resistance observed on ASTs was only evident in two isolates (EC02 and EC03) that were phenotypically resistant to CIP, even though ciprofloxacin ASTs did not correlate with the ARGs identified from WGS results, where 72% of isolates carried fluoroquinolone-resistant genes. This could have resulted from the fact that under laboratory conditions, resistance genes may be present but not expressed; the presence of a resistant gene in a genome doesn’t always indicate its expression [12].

2.6.2. Antibiotic Resistance and Mobile Genetic Elements

Table 3 reveals that above 98% of isolates were positive for the qnrS1 gene, which is a plasmid-mediated quinolone resistance (PMQR) gene. Almost all qnrS1genes were carried on plasmid except for EC12 (contig 41), which was carried on chromosome strain 67 (Accession number CP128443.1). ARGs and MGEs from the E. coli were closely related to the target sequence found in a Genbank database (National Library of Medicine, National Center for Biotechnology Information, USA), with the most hits being plasmids (Table 3). A total of 36% of isolates harboring the qnrS1 gene had quinolone resistance-determining regions (QRDR), which are defined by mutations in the DNA gyrase (gyrA) or in topoisomerase IV (parC) genes known to cause resistance against fluoroquinolones, rendering the antibiotic ineffective. EC03 and EC07 were odd isolates as they exhibited phenotypic resistance that did not correspond with genomic resistance. EC03 had an antibiogram of CIP-TET-SXT-AMP, and EC07 had TET-SXT-AMP, but both harbored the fosA3 resistance gene and had a fluoroquinolone point mutation in gyrA S83L. EC25 exhibited TET resistance phenotypically, and aph(6)-Id, aph(3'')-Ib, tet(B) genomically, but had fluoroquinolone point mutation gyrA (S83L) and a novel parC (K665R).

The study investigated the phenotypic and genomic characteristics of antibiotic-resistant E. coli from poultry litter and wastewater from an intensive poultry production system. No presumptive E. coli was isolated on all floor swabs sampled after disinfection, indicating an effective sanitation method and less risk of transfer of antibiotic-resistant E. coli strains from one poultry flock to the next production cycle.

2.6.3. Antibiotic Susceptibility Profile

To better understand resistance trends in poultry-associated E. coli, we examined the patterns of antibiotic resistance observed across multiple sampling cycles. The resistance patterns remained largely consistent across the cycles, with all E. coli isolates exhibiting relatively high levels of resistance to three antibiotics, i.e., tetracycline (highest prevalence), followed by ampicillin and trimethoprim-sulfamethoxazole. In the December samples, only two isolates showed resistance to ciprofloxacin EC02 and EC03 (Table 1). The findings align with a study by Phiri et al. (10) in Zambia, an analysis of broiler litter, cloacal swabs, and carcass swabs collected from poultry farms, abattoirs, and open markets across seven districts in Zambia was conducted to assess the prevalence and antibiotic resistance profiles of Salmonella and E. coli. The E. coli isolates showed a high level of resistance to ampicillin (68%), tetracycline (81%), trimethoprim/sulfamethoxazole (65%), and ciprofloxacin, being the least (21%). These results are in contrast to the results found by McIver et al. [11] in a study conducted in a similar region on a commercial poultry farm in South Africa. The study assessed the antibiotic-resistant E. coli obtained from litter, feces, wastewater from the chicken house, truck, crates used in transportation, as well as final meat products for consumption. The AST results demonstrated a low level of antibiotic resistance to AMP, TET, and STX. In comparison, in this study, the low level of fluoroquinolone resistance observed on ASTs was only evident in two isolates (EC02 and EC03) that were phenotypically resistant to CIP, even though ciprofloxacin ASTs did not correlate with the ARGs identified from WGS results, where 72% of isolates carried fluoroquinolone-resistant genes. This could have resulted from the fact that under laboratory conditions, resistance genes may be present but not expressed; the presence of a resistant gene in a genome doesn’t always indicate its expression [12].

2.6.4. Antibiotic Resistance and Mobile Genetic Elements

Table 3 reveals that above 98% of isolates were positive for the qnrS1 gene, which is a plasmid-mediated quinolone resistance (PMQR) gene. Almost all qnrS1genes were carried on plasmid except for EC12 (contig 41), which was carried on chromosome strain 67 (Accession number CP128443.1). ARGs and MGEs from the E. coli were closely related to the target sequence found in a Genbank database (National Library of Medicine, National Center for Biotechnology Information, USA), with the most hits being plasmids (Table 3). A total of 36% of isolates harboring the qnrS1 gene had quinolone resistance-determining regions (QRDR), which are defined by mutations in the DNA gyrase (gyrA) or in topoisomerase IV (parC) genes known to cause resistance against fluoroquinolones, rendering the antibiotic ineffective. EC03 and EC07 were odd isolates as they exhibited phenotypic resistance that did not correspond with genomic resistance. EC03 had an antibiogram of CIP-TET-SXT-AMP, and EC07 had TET-SXT-AMP, but both harbored fosA3 resistance gene and had a fluoroquinolone point mutation gyrA S83L. EC25 exhibited TET resistance phenotypically, and aph(6)-Id, aph(3'')-Ib, tet(B) genomically, but had fluoroquinolone point mutation gyrA (S83L) and a novel parC (K665R).

The isolates' resistant profile to fluoroquinolones had no correlation between phenotypic and genomic results, which might have a novel mechanism or are carrying resistance genes that have not yet been annotated in the database [12]. A large number (76%) of isolates conferring resistance to fluoroquinolones was alarming as it might render ciprofloxacin ineffective in the future.

Table 3 also demonstrated a wide range of ARGs associated with MGEs, indicating active and varied horizontal gene transfer among E. coli strains within the poultry farm. The MGEs identified included class 1 integrons (intI1), transposons (Tn3, TnAs1), and ISs (IS, ISKpn19). Most ARGs were carried, some associated with one or more MGEs. five ARGs, dfrA, CmlA1, ANT(3'')-Ia, AadA5 and sul3 were carried on Intl1. Most dfrA and AadA resistant genes carried on intl1 were also associated with TnAs1, whereas ANT(3'')-Ia was either carried on intl1 alone or in association with TnAs1. This was observed on EC01, EC08, EC10, EC12, EC13, EC24, and EC27. DfrA gene from EC21 and EC22 was carried on intl1 only. CmlA1 and sul3 were carried on intl1 and associated with IS256 as referred to EC05, EC18, and EC20. Among the resistant genes harbored by EC02, EC05, and EC20 was QacL. This gene encodes proteins found in the Qac (quaternary ammonium compound) efflux pump family, which contributes to antibiotic resistance. EC05 and EC18 also harbored esX (ESAT-6 secretion system) resistant gene, which confers resistance to macrolides and is known as a virulence factor in Mycobacterium tuberculosis [13]. The esX resistant gene is also associated with the opportunistic M. avium complex (MAC), which causes a chronic infection of avian tuberculosis in birds. Even though modern husbandry has decreased the incidence of avian tuberculosis within the commercial poultry industry, there have been sporadic outbreaks reported in commercial poultry [14]. Another study conducted by Ogundare et al. [15] in South Africa, Gauteng and Limpopo provinces, aimed to determine the virulence profiles and AMR genes of zoonotic APEC, focusing on foodborne EHEC isolated from close human contacts, poultry, swine, and environmental water samples collected from abattoirs and poultry farms, also identified esX gene on their isolates. This might indicate horizontal gene transfer.

2.6.5. Genetic Relatedness

The phylogenomic analysis from our study revealed that several isolates were closely related to poultry E. coli isolates found in Cameroon, Ghana, Nigeria, and Tanzania, and environmental E. coli strains previously identified in South Africa and Ghana (Figure 1). Most of our isolates were closely related to poultry isolates from Ghana and Nigeria, whereas a few exhibited genetic relatedness to isolates circulating in Tanzania. Isolates from our study such as EC15 (562.162049), EC18 (562.162047), had MLST (ST48) as similar to poultry isolates found in Ghana (562.108415), (562.108423), and Cameroon (562.164374), whereas ST155 was shared by isolates EC08 (562.162055), EC23 (562.162040), and isolates from Nigeria poultry (562.110807 and 562.110808). ST48 is commonly associated with foodborne diseases, mostly prevalent in poultry meat [16]. ST155 is also prevalent in poultry; strains carrying ST155 often harbor different antibiotic resistance, like tetracycline, sulfonamides, and β-lactams [17]. ST155 (24%) was the most common sequence type in our study, followed by ST1286 (12%) and ST602 (8%). All ST155 isolates carried the tet(A) gene, which aligns with findings by Davies et al. [18], who similarly reported a high prevalence of ST155 and the associated presence of the tet(A) gene in E. coli from poultry in Bangladesh. One Ghanian (562.135076) environmental isolate shared an ST155 with isolates from the current study. Most of our study isolates were more closely related to poultry isolates than to environmental isolates. The genetic linkage of poultry E. coli isolated from different countries indicates a possible interborder crossing of these strains.

3. Materials and Methods

3.1. Ethical Clearance

Ethical approval was obtained from the University of KwaZulu-Natal Animal Research Ethical Committee (AREC) (Reference number AREC00002891/2021). Additionally, a Section 20 A permit for conducting research on animals was secured from the Department of Agriculture, Land Reform and Rural Development (DALRRD) (Reference number 12/11/1/5 (2283AC)).

3.2. Study Site and Population

The study was conducted at an intensive poultry production system located in the uMgungundlovu District of KwaZulu-Natal, South Africa. The farm comprised of 12 chicken houses, each housing over 25,000 broiler chickens. One chicken house was randomly chosen as the study site. Samples were collected at the end of each production cycle over three consecutive cycles during the summer months: November 2023 for cycle one, December 2023 for cycle two, and January 2024 for cycle three.

Each cycle consisted of 35 days of chicken growth, followed by two days of chicken litter removal (samples of the chicken litter were collected before its removal), one day of chicken house washing (wastewater samples collected on the day of the wash), and one day for disinfecting (floor swab samples were collected 24 hours after disinfection).

3.3. Sample Collection

Sampling was carried out at the end of five weeks, after reaching a full growth production cycle. The chicken house was divided into three rows, and five chicken litter samples were collected from each row at approximately 5m intervals, starting from the front and moving to the back of the house. This resulted in 15 chicken litter samples. To collect the faecal samples, a sterile disposable spatula and sterile zip-lock bags were used. The samples were stored on ice and transported to the Antimicrobial Resistance Unit (ARU), University of KwaZulu-Natal, Westville campus, for processing.

After two days, based on the hygiene maintenance program of the production, wastewater from the first wash was collected from the drain that allows water flow from inside the chicken house, before mixing with wastewater from the rest of the farm. Two wastewater samples were collected in a 2-liter sterile bottle at 30-minute time intervals each. Following 24 hours of disinfecting the chicken house, swabs were taken from the floor, including the corners and the drains inside the house. Each sampling point was separated by five steps; the floor swab sampling resulted in 75 swabs, which were then placed in one sterile bottle containing sterile distilled water (dH₂O), making one sample per sampling cycle. Both wastewater and floor swabs followed the same storage and transportation procedure as previously described.

3.4. Isolation and Identification of E. coli

For the chicken litter, 1 gram of solid litter was added to 9mL sterile dH2O, briefly vortexed, and filtered. Wastewater was thoroughly mixed and filtered prior to use, and floor swab samples were placed on a shaker at 125rpm for 2 h before further processing. A one in 100-fold dilution was conducted, where 1mL of each sample was diluted into 100mL of sterile distilled water. Subsequently, the putative presence of E. coli was determined using the Colilert system (IDEXX Ltd, Westbrook, USA) according to the manufacturer’s instructions, with additional modification (1g of chicken litter was mixed with 9mL dH₂O, filtered before using it in the Colilert system). A single Colilert reagent was added to each sample and mixed well before pouring the mixture onto a Colilert tray that was sealed and incubated at 37˚C for 18 – 24h. After incubation, the trays were viewed under a UV light visualiser (Sigma-Aldrich, Germany, UK) at 260 nm. The wells with blue fluorescence indicated putative E. coli presence. A random selection of wells that fluoresced from each tray was streaked onto a selective media, Eosin Methylene Blue agar (EMB) (Merck, Germany, UK) for putative phenotypic identification and incubated at 37˚C for 18 – 24h for each sample type. After incubation, two isolates were sub-cultured from EMB onto Nutrient agar (Oxoid, Hampshire, England) to confirm culture purity and stored in Tryptic Soy Broth (Oxoid, Hampshire, England) with 20% glycerol at -80˚C for further analysis. Each cycle provided 30 isolates from chicken litter, resulting in a total of 90 isolates for three sampling cycles. For wastewater, each sampling cycle resulted in 2 x 2L samples that provided 20 isolates, resulting in 60 isolates for three consecutive sampling cycles. A total of 150 isolates from all samples constituted the final sample size. No E. coli growth was isolated from floor swab samples. E. coli ATCC 25922 (Oxoid, Hampshire, England) was used as a positive control.

3.5. Molecular Confirmation of Isolates

DNA was extracted using the heat lysis method as previously described (Abrar et al., 2019). The extracted DNA was utilised to confirm E. coli using the uidA gene on a QuantiStudio 5 RealTime PCR System (ThermoFisher Scientific, Waltham, MA, USA) using the forward primer 5’-AAAACGGCAAGAAAAAGCAG-3’ and the reverse primer 5’-ACGCGTGGTTAACAGTCTTGC-3’. All primers were acquired from Inqaba Biotechnical Industries (Pty) Ltd, Pretoria, South Africa. Optimized thermal cycling conditions for uidA (β-D-glucuronidase) where initial uracil-DNA glycosylase (UDG) activation at 50℃ for 2 min, activation of Dual-LockTM polymerase at 95℃ for 2 min, denaturation set at 95℃ for 15s, annealing set at 60℃ for 15s, extension at 72℃ for 10s for 35 cycles, followed by a final extension at 72℃ for 5 min. A high-resolution melting curve was generated by ramping up the temperature from 65℃ to 95℃ at a continuous rate of 0.15℃/s. Each reaction included a positive control, E. coli ATCC 25922, as well as a negative template control, which consisted of nuclease-free water replacing the DNA template.

3.6. Antibiotic Susceptibility Testing (AST)

Confirmed E. coli isolates (n=70), 47% were tested against a panel of ten antibiotics using the Kirby-Bauer disc diffusion method on Mueller-Hinton Agar (Oxoid, Hampshire, England), following the Clinical and Laboratory Standards Institute (CLSI) guidelines [19]. The selection of ten antibiotics aimed to determine whether E. coli from the sampling site showed resistance to 3^rd and 4^th generation cephalosporins, carbapenems, and quinolones, classified as “shared class of antibiotics by animals and humans.” (WHO, 2023). The antibiotics tested were ampicillin (AMP, 10μg), ceftriaxone (CRO, 30µg), cefotaxime (CTX, 30µg), cefepime (FEP, 30µg), ceftazidime (CAZ, 30µg) (CIA), ciprofloxacin (CIP, 5µg), gentamicin (GEN, 10µg), meropenem (MEM, 10µg), tetracycline (TET, 30µg), and trimethoprim-sulfamethoxazole (SXT, 25µg). The diameters of the zones of inhibition were measured and interpreted using the breakpoint criteria provided by the CLSI (2020). E. coli ATCC 25922 was used as a positive control.

Out of the 70 confirmed E. coli isolates, 25 were selected for whole-genome sequencing based on unique antibiograms and MDR profiles, along with two additional isolates that exhibited resistance to tetracycline only. The selection process accounted for all sampling site representation and the three sampling cycles.

3.7. Whole Genome Sequencing Analysis And Bioinformatic Analysis.

Whole genome sequencing was conducted at the National Institute for Communicable Diseases (NICD), Johannesburg, South Africa. Genomic DNA from bacterial samples was extracted using the GenElute Bacterial Genomic DNA Kit (Sigma Aldrich, St. Louis, USA) according to the manufacturer's instructions. DNA concentration and purity were assessed at a 260/280 nm absorbance ratio using a Nanodrop 8000 (Thermo Scientific, Waltham, MA, USA). Subsequently, libraries were prepared using the Nextera XT DNA Library Preparation Kit (Illumina, San Diego, CA, USA) and subjected to whole genome sequencing on an Illumina MiSeq platform (Illumina, San Diego, CA USA). Raw sequence reads were quality-trimmed using Sickle v1.33 (https://github.com/najoshi/sickle) and assembled using the SPAdes v3.6.2 genome assembler. The resulting genomes were submitted to GenBank and assigned accession numbers under BioProject PRJNA1183844. Genome annotation and analysis included identification of antibiotic-resistant genes (ARGs) and disinfectant genes using ResFinder v4.6.0 (https://cge.food.dtu.dk/services/ResFinder/). Plasmid replicon typing was obtained via PlasmidFinder v2.1 (https://cge.food.dtu.dk/services/PlasmidFinder/). Multilocus sequence typing (MLST) was performed using the MLST 2.0 database (https://cge.cbs.dtu.dk/services/MLST/). Mutations conferring resistance to fluoroquinolones and Plasmid/chromosomal sequences with the closest nucleotide homology were analysed using BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastSearch). The DNA gyarase (GyrA and GyrB) as well as the DNA topoisomerase IV (parC and parE) genes were analysed on BLAST using the E. coli ATCC 25922 strain as a reference. Finally, genome synteny and mobile genetic elements (MGEs) were examined on the National Center for Biotechnology Information (NCBI) platform (https://www.ncbi.nlm.nih.gov/).

3.8. Phylogenomic Analyses

A phylogenetic analysis was performed to investigate the relationships among E. coli genomes from this study and those reported in South Africa and other African countries from 2015 to 2024. The genomes were downloaded, annotated, and analysed from the Bacterial and Viral Bioinformatics Resource Centre (BV-BRC) (https://www.bv-brc.org/). The comparison included genomes derived from both poultry litter (n = 11) and wastewater (n = 12). The phylogenetic tree was constructed using the maximum likelihood method on BV-BRC. E. coli K-12 (accession: 511145.12) served as a genome reference. Tree annotation, editing, and visualization were carried out using the Interactive Tree of Life (iTOL) platform (https://itol.embl.de/).

4. Conclusions

This study reveals the role played by MGEs in the spread of ARGs within the poultry setting. Identifying ARGs carried by some associated with MGEs indicates a possibility of horizontal gene transfer taking place within the poultry-associated E. coli. Detection of genomic linkage of our isolates with those from other African countries indicates cross-border transmission. These findings underscore the urgent need to implement an antibiotic stewardship program for the entire continent.