Introduction

Ensuring water safety and sustainable supply is a fundamental human right and a key objective of the UN Sustainable Development Goals (SDGs) that is proposed to be attainable by 20301. This objective, however, faces challenges, with global water scarcity being a pervasive threat in various regions2. In addressing water scarcity, surface water has emerged as a significant source for domestic use, drinking, crop irrigation, and recreation, especially in local communities3. Despite its importance, surface water bodies such as streams, rivers, and lakes are continually impacted by anthropogenic activities, resulting in microbial pollution. Anthropogenic influences stemming from inadequate sanitation practices, untreated wastewater, and fecal-contaminated runoff from farms contribute to the pollution of surface waters, thereby posing a severe public health threat by exposing vulnerable populations to waterborne illnesses4,5,6. These illnesses, caused by the consumption or contact with polluted water containing microbial pathogens, pose health risks to both animals and humans7. As a result, safeguarding surface water quality is imperative to prevent waterborne diseases and protect the well-being of communities relying on these water sources.

Over the years, fecal coliforms and Escherichia coli (E. coli) have served as crucial indicators in water quality monitoring, signalling the potential presence of pathogenic organisms8. Although E. coli typically resides symbiotically in the digestive systems of both humans and animals9,10, certain variants can be virulent, contributing to various enteric infections11. Waterborne outbreaks of gastroenteritis attributed to contaminated water sources have been linked with diarrheagenic E. coli strains12. The significance of E. coli in water microbiology as an indicator of fecal pollution and its public health implications depend largely on the antimicrobial susceptibility pattern exhibited by the organism, which dictates the efficacy of available therapeutic interventions8. Understanding these dynamics is crucial for comprehensive water quality assessment and ensuring effective public health measures.

The dissemination of antimicrobial-resistant bacteria (ARB) and antimicrobial resistance genes (ARGs), particularly when carried on mobile elements like phages or plasmids, poses significant challenges to human well-being13,14. ESBLs, often encoded on plasmids, serve as a vital and widely prevalent resistance mechanism, enabling horizontal transfer among various bacterial groups15. Aquatic environments significantly contribute as sources and pathways for the dissemination of ESBL-producing bacteria16. Exposure to these bacteria, such as E. coli, can happen through direct consumption of polluted water, inadvertent ingestion in the course of recreational activities, or the consumption of fresh crops irrigated with surface water containing ESBLs15. The implications for public health necessitate a comprehensive understanding of these dynamics.

Evaluating ESBL-producing E. coli, especially in domestic water sources, is essential to comprehend the potential for transmitting AMR and causing waterborne diseases. ESBL-producing E. coli has been recorded in surface waters across multiple African nations, including Sudan17, Tanzania18, South Africa19, the Democratic Republic of the Congo20, Ghana21, and Nigeria22. Worldwide, the screening of ESBL-producing E. coli stands as a pivotal measure in AMR monitoring23. In Nigeria, poor adherence to proper water safety practices, indiscriminate usage of antibiotics in veterinary medicine and subsequently, the potential of these animal-related usages in the transfer of AMR via fecal pollution to surface water bodies and other environmental reservoirs have been reported severally1,14,22. The primary focus of this study is to examine the AMR patterns exhibited by both ESBL and non-ESBL-producing strains of E. coli in two distinct surface water sources. Understanding these resistance patterns is essential for addressing the global challenge of AMR in water reservoirs and safeguarding public health.

Results

Cell concentration of E. coli from surface water

E. coli counts in Fig. 1, presented as Log10 CFU/100 mL, vary across stations, locations, and seasons. In Ojerame dam water, then mean E. coli counts range from 1.23–2.18 and 1.10–2.14 across the stations in the wet season and dry season, respectively. In Ovokoto spring water, the mean E. coli counts across stations range from 1.25–1.45 in wet season. The mean count as obtained in Ovokoto spring water (Station 1) in the dry season was 1.84 while other stations lacked water. The observed differences across stations, locations, and seasons are statistically significant based on mean values as observed (p < 0.05).

Fig. 1
figure 1

Cell concentration of E. coli from surface water based on sample source, location and seasonal variability. OJWST1: Ojerame wet season station 1, OJWST2: Ojerame wet season station 2, OJWST3: Ojerame wet season station 3, OJWST4: Ojerame wet season station 4, OJDST1: Ojerame dry season station 1, OJDST2: Ojerame dry season station 2, OJDST3: Ojerame dry season station 3, OJDST4: Ojerame dry season station 4, OVWST1: Ovokoto wet season station 1, OVWST2: Ovokoto wet season station 2, OVWST3: Ovokoto wet season station 3, OVWST4: Ovokoto wet season station 4, OVDST1: Ovokoto dry season station 1, o: represents individual data points that fall outside the upper or lower whiskers of the plot; *: signifies extreme outliers.

Full size image

Occurrence of ESBL and non-ESBL producing E. coli from surface water

Among 36 samples from Ovokoto spring water, 58.3% (21/36) tested positive for ESBL-producing E. coli, while in Ojerame dam water, 33.3% (16/48) tested positive for ESBL-producing E. coli. Additionally, 75% (27/36) of Ovokoto spring water samples and 70.8% (34/48) of Ojerame dam water samples were non-ESBL producers. In total, out of 84 samples, 41.7% (35/84) had ESBL-producing E. coli, and 72.6% (61/84) had non-ESBL-producing E. coli. The E. coli isolates were screened and confirmed based on classical and PCR approaches to give 47 isolates and 67 isolates from Ovokoto spring water and Ojerame dam water, respectively. ESBL-producing E. coli (n = 49) were sourced from Ovokoto spring water (n = 21) and Ojerame dam water (n = 28). Non-ESBL-producing E. coli (n = 65) were sourced from Ovokoto spring water (n = 26) and Ojerame dam water (n = 39).

Antimicrobial susceptibility profile of ESBL and non-ESBL producing E. coli from surface water

The antimicrobial susceptibility analysis of the ESBL and non-ESBL-producing E. coli based on the sample source is shown in Table 1. It was observed that all the ESBL-producing E. coli isolates demonstrated complete resistance (100%) to cefotaxime, ceftriaxone, ceftazidime, and ampicillin. Conversely, these isolates exhibited full sensitivity (100%) to ertapenem, imipenem, meropenem, and nitrofurantoin. Similarly, all non-ESBL-producing E. coli isolates were resistant to ampicillin and displayed sensitivity to ertapenem, imipenem, meropenem, and nitrofurantoin.

Table 1 Antimicrobial susceptibility profile of ESBL and non-ESBL producing E. coli from surface water based on sample source.
Full size table

Multiple AMR characterizations of the ESBL and non-ESBL-producing E. coli isolates

ESBL-producing E. coli displayed 14 diverse MAR phenotypes (MAR index: 0.18–0.46) (Table 2). Their MDR profile ranged from resistance to 5 antimicrobials (3 classes) to 10 antimicrobials (6 classes). Dominant MDR phenotypes included CROR-CTXR-CAZR-AMPR-SSSR (67.3%), CROR-CTXR-CAZR-AMPR-AZMR (51.0%), and CROR-CTXR-CAZR-AMPR-TETR (44.9%). Non-ESBL-producing E. coli had 9 MAR phenotypes (MAR index: 0.12–0.5). Their MDR profile spanned resistance from 5 antimicrobials (3 classes) to 11 antimicrobials (8 classes). Predominant MDR phenotypes included AMPR-PTZR-TETR-CHLR-PZPR-AZMR-SSSR-SXTR-AMKR-STRR-CIPR (6.2%), AMPR-PTZR-TETR-CHLR-PZPR-AZMR-SSSR-SXTR (26.2%), and AMPR-PTZR-AMKR-STRR-CIPR (18.5%).

Table 2 Distribution of antimicrobial resistance in ESBL and non-ESBL producing E. coli isolates.
Full size table

Classical oxacillinases, ampC beta-lactamases, carbapenemases/metallo-beta-lactamases and ESBL determinants from ESBL-producing E. coli based on sample sources

The classical oxacillinases (OXA-1 and OXA-47), ampC beta-lactamases (blaCMY−2), carbapenemases/metallo-beta-lactamases (blaNDM−1) and ESBL gene (TEM and CTX) percentages in Fig. 2 reveal their prevalence in ESBL E. coli from water sources. Notably, in Ojerame dam water, blaOXA-1 is 7.1%, blaTEM is 64.2%, blaCTX−M−15 is 71.4%, and blaCTX−M−1 is 85.7%. In Ovokoto spring water, the percentages are blaTEM 52.4%, blaCTX−M−15 85.7%, and blaCTX−M−1 95.2%. Overall ESBL gene proportions in ESBL E. coli include blaOXA-1 4.1%, blaTEM 59.2%, blaCTX−M−15 77.6%, and blaCTX−M−1 89.8%. Others such as blaCMY−2, blaCTX−M−8, blaCTX−M−9, blaCTX−M−2, blaSHV, blaOXA−47 and blaNDM−1 were not detected.

Fig. 2
figure 2

Prevalence of classical oxacillinases, beta-lactamases and carbapenemases/metallo-beta-lactamases genes from ESBL-producing E. coli based on sample source.

Full size image

Multi classical oxacillinases and ESBL resistance genes from ESBL-producing E. coli based on sample source

The distribution of combined classical oxacillinases and ESBL genes in Fig. 3 highlights their prevalence in ESBL-producing E. coli based on the sample source. Notably, in Ojerame dam water, blaOXA-1 + blaCTX−M−1 is 7.1%, while in Ovokoto spring water, blaTEM + blaCTX−M−1 is 23.8%, blaTEM + blaCTX−M−15 is 19.1%, and blaCTX−M−15 + blaCTX−M−1 is 57.1%. Overall, combined classical oxacillinases and ESBL gene proportions are blaOXA-1 + blaCTX-M-1 4.1%, blaTEM + blaCTX-M-1 16.3%, blaTEM + blaCTX-M-15 22.5% and blaCTX-M-15 + blaCTX-M-1 55.1%.

Fig. 3
figure 3

Prevalence of multiple classical oxacillinases and ESBL resistant genes from ESBL E. coli based on sample source.

Full size image

Antimicrobial resistance genes of the E. coli isolates are based on water sources

The distribution of antimicrobial resistance genes in E. coli isolates from Ovokoto Spring and Ojerame Dam waters (Fig. 4) is outlined. Key findings include tetA [Ovokoto 10.6%, Ojerame 3.0%], tetC [Ovokoto 14.9%, Ojerame 13.4%], tetM [Ovokoto 44.7%, Ojerame 34.3%], sul1 [Ovokoto 38.3%, Ojerame 35.8%], sul2 [Ovokoto 25.5%, Ojerame 29.8%], sul3 [Ovokoto 6.4%, Ojerame 11.9%], qnrA [Ovokoto 10.6%, Ojerame 4.5%], qnrB [Ovokoto 4.3%], and qnrS [Ovokoto 14.9%, Ojerame 5.9%]. Mobile genetic elements were also detected in proportions such as class 1 integron [Ovokoto 12.8%, Ojerame 22.4%] and class 2 integron [Ovokoto 6.4%, Ojerame 10.5%]. Cumulatively, ARGs in E. coli isolates obtained from both water sources include tetA 6.1%, tetC 14.0%, tetM 38.6%, sul1 36.8%, sul2 28.1%, sul3 9.6%, qnrA 7.0%, qnrB 1.8%, and qnrS 9.6%. Mobile genetic elements such as class 1 integron and class 2 integron were detected in proportions as 18.4% and 8.8%, respectively. Furthermore, the results of the AST are consistent with the profiles of ARGs, as all the isolates harbouring ARGs demonstrated phenotypic resistance to the tested antimicrobials.

Fig. 4
figure 4

Distribution of antimicrobial-resistance genes and mobile genetic elements in the E. coli isolates based on sample sources.

Full size image

Antimicrobial resistance genes from the E. coli isolates based on ESBL production potentials

The distribution of antimicrobial resistance genes in non-ESBL-producing and ESBL-producing E. coli isolates is illustrated in Fig. 5. Notable findings include tetA [ESBL-producing E. coli 6.1%, non-ESBL producing E. coli 6.2%], tetC [ESBL-producing E. coli 24.5%, non-ESBL producing E. coli 6.2%], tetM [ESBL-producing E. coli 36.7%, non-ESBL producing E. coli 40%], sul1 [ESBL-producing E .coli 57.1%, non-ESBL producing E. coli 21.5%], sul2 [ESBL-producing E. coli 42.9%, non-ESBL producing E. coli 16.9%], sul3 [ESBL-producing E. coli 16.3%, non-ESBL producing E. coli 4.6%], qnrA [ESBL-producing E. coli 10.2%, non-ESBL producing E. coli 4.6%], qnrB [ESBL-producing E. coli 4.1%], and qnrS [ESBL-producing E. coli 12.2%, non-ESBL producing E. coli 7.7%]. Mobile genetic elements were also detected in proportions such as class 1 integron [ESBL-producing E. coli 28.6%, non-ESBL producing E. coli 10.8%], and class 2 integron [ESBL-producing E. coli 14.3%, non-ESBL producing E. coli 4.6%]. The frequency of each resistance gene and mobile genetic elements category is summed up across all isolates in the study. This provides an overall representation of the prevalence of each resistance gene and mobile genetic elements category in the entire sample population of isolates, regardless of whether they are ESBL-producing or non-ESBL-producing E. coli. The chloramphenicol acetyltransferase gene plasmids (cat::pC194, cat::pC223, and cat::pC221) were not detected from isolates in the study.

Fig. 5
figure 5

Antimicrobial-resistance genes and mobile genetic elements of the E. coli isolates based on ESBL production capacity.

Full size image

Discussion

The surge in the prevalence of ESBL-producing E. coli and associated genotypes has triggered significant public health concerns, given their correlation with increased illness rates, limited therapeutic options, prolonged hospital stays, and elevated fatality rates. This study investigates the cell concentration of E. coli, ESBL-producing E. coli, and non-ESBL-producing E. coli in two surface waters in Nigeria. Comparable studies on surface water in Nigeria reported higher E. coli concentrations (3.00 to 5.75 Log10 CFU/100 mL)24, while Ethiopia25 and Germany26 reported ranges of 2.4 to 3.7 Log CFU/100 mL and 1.77 to 4.43 Log10 CFU/100 mL, respectively. A significant difference was observed across stations, locations, and seasons (p < 0.05). This observation aligns with prior studies reporting higher E. coli counts in wet seasons compared to dry seasons3,25, a fact emphasizing the seasonal dynamics of microbial populations in surface waters. However, it must be noted that while samples from wet and dry seasons were seamlessly obtained from all four stations in Ojerame surface water all year round, the same was not possible for Ovokoto surface water as three out of four of the stations dried up during the dry season. Since the sampling from the Ovokoto surface water is very weak, no conclusion can be drawn from the presented results.

During wet seasons, surface waters often experience an elevated microbial contamination level attributed to increased rainfall, leading to runoff from agricultural and domestic wastes, as well as fecal matter due to inadequate drainage and waste management systems6. The detection of E. coli in surface water signals the presence of fecal contamination. This study observed higher E. coli cell concentration at Ojerame Station 3, which could be attributed to extensive agricultural activities in that area. However, the recorded E. coli concentrations remained below the recommended safety thresholds for irrigation (2.35 Log10 CFU/100 mL) based on FAO and WHO recommended limit27. Similarly, for bathing, body contact, and recreation, the concentrations were below the US Environmental Protection Agency’s (USEPA) suggested safe criterion of 2.37 Log10 CFU/100 mL28. Nevertheless, FAO and WHO emphasized that E. coli should be zero CFU/100 mL in drinking water and 126 CFU/100 mL, which is approximately 2.10 Log10 CFU/100 mL in water for domestic and recreational water purposes4. In view of this, the water is deemed unsafe for drinking, as per WHO guidelines, which mandate a zero CFU/100 mL standard for microbial quality in drinking water29. Furthermore, previous research has also indicated that the absence of E. coli in water samples using conventional culture-based methods is not entirely conclusive as there are still tendencies of E. coli in a viable but not culturable state8. Hence, advanced methods with improved analytical sensitivity, which involve a larger volume of samples, such as dead-end ultrafiltration and droplet digital PCR, are recommended in the detection of water pathogens due to their better information pathogen-specific risk assessment and management in the assessment of microbial water quality8. This indicates that a larger volume of samples could increase the precision of etiological agents’ surveillance in water and curtailing waterborne outbreaks. Additionally, despite the low concentration of E. coli in water sources, potential health risks also persist in their usage for domestic and recreational purposes, particularly due to their association with risk factors such as ESBL.

ESBL-producing E. coli isolates and non-ESBL-producing E. coli isolates were recovered from water samples in this study. Comparable studies on ESBL-producing E. coli reported detection rates ranging from 1.7%–98% across various African countries, including Sudan (40%) [North Eastern Africa]17, Tanzania (1.7%) [Eastern Africa]18, Ghana (98%) [Western Africa]21 and Nigeria (22.7%) [Western Africa]22, which aligns with the findings from Ojerame and Ovokoto water in this study. ESBL-producing E. coli occurrence appears linked to a high concentration of E. coli, but the rates varied across sites and seasons. In Ovokoto Spring, the occurrence of ESBL-producing E. coli was higher than in Ojerame Dam water. This could be linked to fecal pollution arising from diverse animals girdling around the water source alive; the animals were brought in for ritual purposes. This aligns with the notion that animal feces contribute considerably to ESBL-producing E. coli in surface water, especially in rural regions characterized by a mix of human and animal origins3.

Our study revealed comparable resistance levels in both ESBL-producing E. coli and non-ESBL-producing E. coli to ampicillin, alongside shared sensitivity to ertapenem, imipenem, meropenem, and nitrofurantoin, consistent with previous study11. The observed ampicillin resistance may stem from its indiscriminate use compared to cephalosporins of later generations. Conversely, the high susceptibility of E. coli to ertapenem, imipenem, meropenem, and nitrofurantoin underscores their effectiveness in treating E. coli infections, aligning with prior reports30. In further comparison with another report, our study of E. coli isolates demonstrated varying resistance to multiple antimicrobial agents that were tested31. The proximity of cattle, human communities, and agricultural activities near water bodies likely contributes to increased bacterial resistance in surface water, either directly or indirectly. The elevated resistance levels to commonly used antimicrobials suggest potential antimicrobial misuse or overuse in both human medicine and veterinary practices, as earlier emphasized32.

The prevalence of AMR E. coli varies in surface waters, influenced by factors such as usage, region, fecal contamination levels, sampling time, and seasons. In our investigation, the recognition of MDR E. coli was expected, considering the existence of remains from sacrificial animals and adjacent agricultural areas. Agricultural runoff, carrying fertilizers into surface waters, likely contributed to this resistance pattern14. The rise of AMR in bacteria presents a substantial challenge in both human and veterinary medicine, given the widespread use of these drugs for treating animals and humans10. The excessive and indiscriminate application of antimicrobials in animal healthcare and the production of animal feed is suspected of contributing to the escalation of AMR in pathogenic bacteria, as noted in WHO reports33.

In this investigation, E. coli exhibited the highest resistance towards piperacillin, contrasting a prior study on surface water where the highest resistance was recorded in tetracycline30. This discrepancy could stem from regional variations in the choice of antimicrobials despite their common availability. Penicillin and tetracycline, widely used in developing countries for treating respiratory issues, diarrhea in humans, and infections like mastitis in cattle, may contribute to this variation34. Notably, streptomycin displayed low resistance, which can be potentially attributed to its limited use and intravenous administration, alleviating concerns30. The study revealed that both non-ESBL-producing and ESBL-producing E. coli were MDR to a minimum of five antimicrobials across three antimicrobial classes and a maximum of ten antimicrobials within six different antimicrobial classes.

The findings of this study revealed that all MDR ESBL-producing E. coli exhibited resistance to ceftriaxone, cefotaxime, ceftazidime, ampicillin, and aztreonam. This aligns with the outcome of previous research where most ESBL-producing strains of E. coli exhibited resistance to three or more commonly used antimicrobials35. Comparable results from surface water in Nigeria demonstrated resistance to a range of antimicrobials, spanning from two to seven agents, encompassing tetracycline, amoxicillin, and cotrimoxazole36. Additionally, the significantly high level of resistance displayed by the non-ESBL producing E. coli could have been influenced by elevated production of β-lactamase an enzyme that neutralizes antimicrobial activity via mechanisms other than ESBLs, such as the AmpC β-lactamases and carbapenem resistance mechanism37. Additionally, 96.7% of isolates from drinking water samples in Hyderabad, India, exhibited resistance to a range of two to six antimicrobials38, highlighting varied resistance patterns to multiple antimicrobial agents.

The surge in MDR bacteria poses a significant concern, given their plasmid-carried genes that exhibit high transferability, as emphasized13. The aquatic environment is witnessing an escalation in MDR bacteria, imposing selective pressures on indigenous bacterial communities, as previously noted39. Notably, the sampled stations were surrounded by various anthropogenic activities, indicating a potential high-risk contamination source for water. The widespread AMR in the environment is underscored by MAR indices surpassing the critical threshold of 0.2. MAR index threshold of 0.2 could discern contamination levels, distinguishing between low and high-risk sources40. However, MAR can also result from the accumulation of diverse resistance mechanisms over time in an organism41. The elevated MAR indices suggest that non-ESBL-producing and ESBL-producing E. coli originate from pollution sources with high contamination risk, where antimicrobials are frequently used, particularly in veterinary medicine.

In this investigation, blaCTX-M-1 exhibited a high prevalence of 89%, detected in 44 out of 49 isolates, while blaCTX-M-9, blaCTX-M-2, blaCTX-M-8, blaSHV, blaOXA-47, blaNM-1, blaOXA-1, blaCMY-2 were not identified. Notably, blaCTX-M-1 is frequently found in livestock and animal-related environments, aligning with its predominant role in this study42. Although blaCTX-M-15 is commonly associated with human isolates, previous studies have reported the detection of blaCTX-M-15-encoding E. coli in animals43. The abundance of blaCTX-M-15 in water sources exposed to animal feces44 further supports our results, suggesting a contamination route likely linked to animal sources for the studied water bodies. Corroborating our study, investigations in Bejaia, Algeria45 and Tehran, Iran46, have previously reported the presence of CTX-M genes in drinking and fresh waters, emphasizing the global concern of ESBL contamination in water sources.

The study revealed a notable prevalence of phenotypic ARGs in surface waters, with sul1 emerging as the most prevalent gene. This finding contrasts with a comparable study, where sulII was the predominant occurrence30. The identification of these resistance genes suggests the widespread misuse or overuse of corresponding antimicrobials in the community. Notably, sul1 and sul1I have been previously identified in isolates from the English Channel, North Sea Sectors47, and shrimp ponds and wastewater in North Vietnam48. The detection of sulII and sulIII in our study aligns with similar findings reported in diverse locations49. In aquatic settings such as rivers, wastewater, and surface water, sulII and sulIII have been recognized as the most prevalent sul genes contributing to sulfonamide resistance50.

In the analysis of tetracycline resistance genes, tetC, tetA, and tetM were identified in both ESBL and non-ESBL-producing E. coli isolated from the two surface water sources, with tetM emerging as the most prevalent across all samples. This aligns with the elevated rates of tetracycline resistance genes detected in urban recreational lakes in China51. The observed variations in the detected tet genes between regions could be influenced by diverse tetracycline antimicrobial usage and the bacterial species present in those regions. Concerning quinolone resistance genes, qnrS, qnrB, and qnrA were detected, with qnrS exhibiting the highest prevalence at 9.6%, although relatively lower compared to other study reports. The plasmid-mediated quinolone resistance nature of the qnrA gene and its dissemination potential among Enterobacteriaceae has been highlighted to be a notable factor of influence in quinolone resistance52. The presence of these genes is probably linked to agricultural practices near the surface water, considering the widespread utilization of quinolones in contemporary agriculture and veterinary medicine.

The identification of class 2 and class 1 integrons in this study resonates with previous reports of E. coli isolates from India53. Notably, class 2 and 3 integron was absent in a previous study39, with only class 1 integron detected39 which doesn’t align with findings in our study. E. coli isolates containing class 1 integrons have been linked with a remarkably increased prospect of MDR compared to those without these integrons39. Additionally, class 1 integrons often carry genes conferring resistance to heavy metals and disinfectants, suggesting that their presence in the study area may be associated with the frequent use of disinfectants in daily activities like washing and bathing. The horizontal transfer of integron genes between strains of various origins is an essential mechanism that could influence the extensive spread of AMR among pathogenic E. coli and other waterborne bacterial pathogens in the region39.

Conclusion

This study unveils the AMR profiles of ESBL and non-ESBL-producing E. coli isolated from two distinct surface waters in Akoko Edo, Nigeria. Significantly, it marks the initial documentation of AMR patterns, MDR, and resistance genes in ESBL and non-ESBL-producing E. coli strains within this community. The presence of ESBL-producing E. coli strains in the surface waters signals potential hazards, rendering them unsafe for consumption and raising concerns about domestic and recreational usage. The widespread distribution of MDR genes in these waters underscores the substantial risks associated with their utilization and highlights the potential role of surface water in the dissemination of AMR. Given the implications for public health posed by MDR bacteria, urgent and effective measures are imperative to mitigate their spread. Additionally, fostering regular awareness campaigns on safeguarding aquatic environments, coupled with adherence to guidelines for pollution prevention, becomes essential in protecting water quality and public well-being.

Materials and methods

Description of the study site and sample collection

The study area includes two surface water bodies, Ovokoto Spring and Ojerame Dam, in Akoko-Edo, Edo-North Senatorial District, Edo State, Nigeria. The water samples were collected from four (4) different sample collection points, namely Station 1, Station 2, Station 3 and Station 4 in each of the surface water bodies, with station 1 representing the source while stations 2–4 representing different points along the downstream. Using sterile 1-L plastic containers, grab samples of water were collected monthly at the different sampling points at a depth of about 30 cm, kept on ice, and analyzed at the Applied Microbial Process and Environmental Health Research laboratory within 5–8 h of collection due to proximity. The Ovokoto Spring holds spiritual significance and serves as a crucial drinking water source for the community. Access is restricted to the source (Station 1), which the locals directly consume as they believe in its spiritual-medicinal properties, while the other three stations serve various domestic purposes. During the dry season, stations 2, 3, and 4 may dry up, but in the wet season, all stations contain water. Ojerame Dam is a construction made across a river named Onyami. Its height is 3.9 m, and it has a water storage capacity of 900,000 gallons. The locals use the dam to supply water for domestic purposes, such as water sources the residents use for drinking, washing, recreational activities, and agricultural purposes activities, which include crop irrigation and drinking water for animals. The Ojerame Dam Station 3 is primarily used for palm oil milling at a nearby processing plant, with the plants’ waste discharge visibly impacting the water quality. However, both surface water bodies were also exposed to several other sources of pollution, resulting in direct defecation by animals such as birds and cattle, runoff from agricultural soils during rainfall and diverse anthropogenic activities. The samples were collected monthly over 12 months (December 2018 to November 2019) and categorized by seasons—dry season (November to February) and wet season (March to October).

Isolation of E. coli and evaluation of their cell concentration

The isolation and quantification of E. coli were carried out using the membrane filtration method. In brief, water samples amounting to 100 mL each were filtered through a 0.45 μm sterilized filter membrane, and the filtered material was transferred to chromocult agar (Merck Darmstadt, Germany). After 24 h of incubation at 37 °C, colonies displaying purple/violet coloration on chromocult agar were counted as E. coli (expressed in CFU/100 mL). Upon obtaining purple/violet colonies, 1 to 3 well-isolated colonies were purified aseptically on nutrient agar medium (Lab M, Lancashire, UK) through the streak-plate technique and subsequently incubated at 37 °C for 24 h. The purity of E. coli samples was confirmed through the Analytical Profile Index (API) 20E tests, following the manufacturer’s guidelines. Data analysis utilized the API database (version 4.1) and APIweb™ identification software. Following purification and confirmation, E. coli specimens were preserved on nutrient agar slants at 4 °C until needed for subsequent research purposes. E. coli ATCC 25,922 served as a control strain.

DNA extraction and polymerase chain reaction amplification E. coli

Isolated and purified E. coli, subjected to DNA extraction, were revived by placing them in tryptone soy broth (TSB) (LAB M, Lancashire, United Kingdom) and incubating at 37 °C for 18 h. DNA extraction was conducted using the PeqGold Bacterial DNA kit (Peqlab Biotechnologie GmbH, Germany) based on the manufacturer’s instructions. Quality and quantity of total genomic DNA were assessed using a NanoDrop TM 1000 Spectrophotometer (Thermo Fischer Scientific, US) and agarose gel electrophoresis54. The uidA gene was amplified using the E. coli-specific primer pair (Supplementary Table S1). E. coli ATCC 25,922 served as a control strain. PCR reactions (50 μL) included 10 μL gDNA, 2.5 μL R primer (adjusted to 10 pmol/μL), 2.5 μL F primer (adjusted to 10 pmol/μL), 5 μL PCR buffer with MgCl2, 23.7 μL nuclease-free water, and 6 μL dNTP mix, 0.3 μL Taq polymerase55. PCR was carried out by employing the Peltier-Based Thermal Cycler (Zhengzhou Mingyi Instrument, China). Gel electrophoresis employed a 1.0% agarose gel with GelRed, and the gel ran for one hour at 100 V DC voltage.

Detection of extended-spectrum beta-lactamase (ESBL) producing E. coli

The detection of ESBL-producing E. coli using CHROMagar ESBL media (75,006 Paris, France) followed established protocols from prior studies56. The E. coli isolates were streaked on the media plates and incubated at 37 °C for 18–24 h. The detection of ESBL production was determined by observing characteristic colonial morphology and growth on the culture media. Dark pink to reddish colonies on CHROMagar ESBL plates signified the probable presence of ESBL-producing E. coli. Colonies from positive samples were selected and subjected to ESBL phenotypic confirmation using the double-disk test57. The detection of ESBL production was carried out using the double disc synergy test (DDST) protocol with ceftazidime-clavulanate (30/10 μg) and ceftazidime (30 μg), as well as cefotaxime-clavulanate (30/10 μg) and cefotaxime (30 μg). After incubating at 37 °C for 18 h, we evaluated ESBL production based on CLSI guidelines57. We identified it by a ≥ 5-mm increment in the inhibition diameter when supplemented with clavulanate compared to the zone diameter of the agent test assessed alone. E. coli ATCC 25,922 served as a control strain.

Antimicrobial susceptibility profile of ESBL and non-ESBL producing E. coli from surface water

Antimicrobial susceptibility assessment for both ESBL E. coli and non-ESBL E. coli was assayed using the Kirby-Bauer disk diffusion method, following Clinical and Laboratory Standards Institute (CLSI) guidelines57. A total of 22 antimicrobial agents were assayed against ESBL and non-ESBL-producing E. coli isolates using antimicrobial disks from Mast Diagnostics, UK. These included various classes such as penicillins (ampicillin 10 μg, piperacillin 100 μg), β-lactam combination agents (ceftolozane-Tazobactam 30/10 μg, piperacillin-Tazobactam 100/10 μg), monobactams (aztreonam 30 μg), tetracyclines (tetracycline 30 μg), carbapenems (meropenem 10 μg, ertapenem 10 μg, imipenem 10 μg), quinolones and fluoroquinolones (levofloxacin 5 μg, ciprofloxacin 5 μg, nalidixic acid 30 μg), nitrofurans (nitrofurantoin 300 μg) and phenicols (chloramphenicol 30 μg), aminoglycosides (gentamicin 10 μg, amikacin 30 μg, streptomycin 10 μg), folate pathway antagonists (trimethoprim-sulfamethoxazole 1.25/23.75 μg, sulfisoxazole 300 μg), cephems (cefotaxime 30 μg, ceftazidime 30 μg, ceftriaxone 30 μg). E. coli ATCC 25,922 served as a control strain.

Multiple AMR characterizations of the ESBL and non-ESBL-producing E. coli

ESBL and non-ESBL-producing E. coli strains manifesting resistance to a minimum of three antimicrobial classes were categorized as multidrug-resistant (MDR). The Multiple Antimicrobial Resistance (MAR) profile of these E. coli types was assessed based on their resistance to three or more antimicrobial classes from a total of fourteen classes. The MAR index for both ESBL and non–ESBL E. coli was calculated using the formula:

$${\text{MAR}}\;{\text{index}} = \frac{b}{c}$$

Here, ‘b’ represents the cumulative count of antimicrobials to which ESBL and non–ESBL E. coli exhibited resistance. At the same time, ‘c’ indicates the total number of antimicrobials utilized against the bacterial species. Previous reports have indicated that intensive antimicrobial usage in the studied area is one of the factors associated with MARI exceeding 0.2, which suggests a high-risk environment for the proliferation of AMR40.

Antimicrobial resistance gene profiling

Genomic DNA from the bacterial lysates was employed as the DNA source in a PCR for the identification of antimicrobial resistance genes. Detection targeted genetic elements associated with ESBL, encompassing blaCTX−M−15, blaSHV, blaTEM, and all the CTX-M groups (blaCTX−M−2−group, blaCTX−M−9−group, blaCTX−M−1−group, blaCTX−M−14−group, blaCTX−M−8−group), including the blaOXA−1−group (classical oxacillinases). Additionally, PCR amplification was carried out for blaOXA−47, blaNDM−1 (classical oxacillinases and carbapenemases/metallo-beta-lactamases), and blaCMY−2 genes (AmpC beta-lactamases) according to established conditions58. Resistance determinants for tetracycline [tetB, tetA, tetM], sulfonamides [sul3, sul2, sul1], aminoglycosides [aacC(3)-1, ant(4')-Ia], quinolones [qnrB, qnrA, qnrS, qnrC], and chloramphenicol [cat3, cat2, cat1] were assessed, as reported by Sáenz et al. (2004). The investigation also included the determination of intI2 and intI1 genes, which encode the class 1 and class 2 integrases, following the methodology previously outlined59. Thermocycling conditions and specific primers are detailed and described in supplementary Table 1 (S1 Table).

Statistical analysis

Microsoft Excel and IBM SPSS were employed for data analysis. Descriptive statistics, encompassing mean and standard deviations, were utilized to summarize the data. A one-way ANOVA was subsequently employed to compare multiple variables, followed by Duncan’s multiple range test. Significance was determined at p-values below 0.05.