Comprehensive insight on multidrug resistance and virulence genes of ESBL-producing E. coli from different surface water sources in Bangladesh

Antimicrobial resistance (AMR) is an emerging concern worldwide from clinical and environmental aspects (Abayneh et al. 2023; Ramatla et al. 2023). Irrational use of antibiotics in accordance with the impact of human activities linked to microbial ecosystems have facilitated the evolution and spreading of antibiotic resistance in the environment (Byrne et al. 2019; Bottery et al. 2021). Wastewater from industries, agricultural runoff, and improper disposal in the environment engender selective pressure, which favours the proliferation of resistant microbes (Manyi-Loh et al. 2018; Polianciuc et al. 2020). Resistance of environmental microbes to antibiotics accounts for challenges from two perspectives (Kasanga et al. 2023). From the clinical point of view, these resistance mechanisms have the potential to undermine the efficiency of antimicrobial therapies hindering the alternatives available for managing bacterial infections (Byrne et al. 2019; Abayneh et al. 2023; Darby et al. 2023). In regard to the environment, human health could be directly exposed to the resistant microbes through contaminated water supplies, food, and air, therefore increasing the risk of infection with untreatable infections (Berglund 2015; Polianciuc et al. 2020). The World Health Organization (WHO) has classified extended-spectrum β-lactamases (ESBLs) in Enterobacteriaceae as the major leading cause of AMR, necessitating the development of new antibiotics (WHO 2017). These bacteria are leading to increased usage of costly antibiotics, longer hospital stays, higher morbidity, death, and healthcare expenses. ESBLs are plasmid-mediated enzymes that break down the structure of the beta-lactam group of antibiotics such as penicillin, aztreonam, and extended-spectrum cephalosporins. Genetic determinants for producing these ESBLs are the bla_SHV, bla_TEM, bla_OXA, and bla_CTX-M genes (Ramatla et al. 2023). Escherichia coli (E. coli), from the Enterobacteriaceae family, is a commensal inhabitant in the intestine of humans and warm-blooded animals. The environmental dissemination of E. coli is primarily linked to faecal contamination from humans and animals. Therefore, E. coli is also widely utilized as an indicator organism to assess faecal contamination in environmental waters, signifying the presence of potentially harmful pathogens (Kasanga et al. 2023). Although most of the E. coli strains are commensals and benign, particular variants can pose a significant threat to human health and the environment through the acquisition of antibiotic resistance and virulence properties (Antonelli et al. 2021). AMR in environmental E. coli evolve from the omnipresent use of antibiotics in human and veterinary treatments, poultry feeds, and agriculture practices (Franz et al. 2015; Pormohammad et al. 2019). In addition, there are six types of E. coli based on their virulence properties, which are named as Enteropathogenic (EPEC), Enteroinvasive (EIEC), Shiga toxin-producing (STEC), Enteroaggregative (EAEC), Enterotoxigenic (ETEC), and diffusely adhering E. coli (DAEC). Surveillance on the presence of ESBL-producing and virulent E. coli in the environment can lessen the risk for public health in terms of drug resistance and pathogenicity (Franz et al. 2015; Mou et al. 2023).

Bangladesh, a densely populated and agricultural country in South Asia, is facing challenges due to the surge of antibiotic resistance in both clinical and environmental settings, particularly in various water bodies (Franz et al. 2015; Hoque et al. 2020). The country's reliance on water for domestic, agricultural, and industrial needs, combined with the widespread use of antibiotics in healthcare, poultry feeds, and agriculture, has contributed to the development of AMR within its environmental matrices (Bilal et al. 2023; Mou et al. 2023). Water bodies in Bangladesh have evolved into substantial repositories for antibiotic residues and resistant bacteria, as a result of contamination caused by different industrial and human activities (Islam et al. 2012; Ahmed et al. 2019). Previous studies reported the prevalence of E. coli with multidrug resistance (MDR) and ESBL phenotypes in water sources of Bangladesh, most of which were drinking water and river water samples (Haque et al. 2014; Rashid et al. 2015; Parvez et al. 2017; Islam et al. 2023). In one of our previous studies, we found 100% of MDR in ESBL-producing E. coli isolated from surface water adjacent to the pharmaceutical industries in Bangladesh (Mou et al. 2023). The concurrent presence of MDR and ESBL production in E. coli from contaminated water sources poses a drastic challenge from both water quality control and public health perspectives. Monitoring and understanding the dynamics of the antibiotic resistance in different water sources are therefore crucial for developing strategies to mitigate the environmental dissemination of resistant bacteria. Therefore, we aimed to compare the antibiotic resistance pattern, the genes associated with ESBL resistance, and pathogenic types of E. coli in surface water sources proximate to the pharmaceutical facilities, agricultural fields, poultry farms, and textile industries of Bangladesh.

Savar, which lies between 23°44′ N and 23°12′ N latitude and between 90°11′ E and 90°22′ E longitude, is swiftly developing hubs of urbanization and industrial activities in Bangladesh. More than 1,500 industries are presently operational in two major export-processing zones in Savar, which include a variety of pharmaceutical, textiles, leather processing, and dyeing industries (Hasan et al. 2022). Besides, innumerable poultry farms have mushroomed in this area over the last several years (Haque et al. 2021). Agricultural land comprises 25% of Savar, which has significantly diminished over the years due to the extensive urbanization and industrialization (Ahmed et al. 2022). Savar is surrounded by numerous water bodies that receive wastewater from these adjacent industries, poultry farms, and agricultural lands. The control samples we collected are from a residential area, which is a public university campus in Bangladesh. There are a number of natural lakes here, from which we have selected two samples as control samples. These lakes are not connected or exposed to any polluting activities, as per our knowledge and the spot observation. Moreover, these water sources are used for pisciculture; therefore, the residents are careful not to dispose any waste there.

The collected samples were subjected to bacteriological quality analysis using conventional culture methods. In brief, the samples were serially diluted up to 10⁻⁴ using normal saline (0.85%) and 1 mL of each dilution was then spread onto non-selective nutrient agar (NA) (Oxoid, UK), selective MacConkey agar (HIMEDIA), and ESBL CHROM agar supplemented with ceftriaxone antibiotic (0.57 mg/mL) for total bacteria, total Gram-negative bacteria, and total ESBL-producing bacterial count, respectively. All the plates were kept at 37 °C for 18–24 h incubation. The appearance of characteristic colonies on each plate was recorded and enumerated. Distinct presumptive colonies of E. coli were picked from the MacConkey plates and ESBL CHROM agar plates, purified further in the nutrient agar plates, and the pure cultures were taken for further preservation and analysis. All the bacterial cultures were stored in nutrient broth at −20 °C with 50% glycerol.

The presumptively identified E. coli isolates were patched onto eosin methylene blue (EMB) agar for further confirmation. From the EMB agar plates, isolates with green metallic sheen were selected for molecular amplification of the E. coli-specific uidA gene by polymerase chain reaction (PCR) (Cleuziat & Robert-Baudouy 1990). The modified boiling DNA method was performed to extract the genomic DNA from the bacterial isolates (Sultana et al. 2017). Shortly, a single pure colony of the isolates was inoculated into 5 mL of nutrient broth and incubated at 37 °C for 24 h. After that 1 mL of culture was transferred to 1.5-mL eppendorf and centrifuged for 10 min at 12,000 rpm. Pellets were collected and washed with sterile deionized water, re-centrifuged, and suspended in 200 μL of PCR grade water. Next, the tubes were kept at 100 °C for 10 min for boiling and then they were immediately placed on ice for 10 min. Finally, the tubes were centrifuged at 10,000 rpm for 10 min and the supernatant (100 μL) was collected into 1.5 mL eppendorf.

PCR was employed to detect the E. coli-specific uidA gene using the primer sets UAL-754 and UAR-900 (Cleuziat & Robert-Baudouy 1990; Titilawo et al. 2015; Table 1). PCR amplification was carried out with 25 μL reaction mixture consisting of 10.5 μL nuclease-free water, 12.5 μL G2 green master mix, 0.5 μL forward primer, 0.5 μL reverse primer, and 1 μL of DNA template. The final concentration of the primers in the reaction was 10 pmol. The following reaction conditions were maintained: denaturation step at 94 °C for 2 min, followed by 25 cycles of denaturation at 94 °C for 1 min, annealing at 58 °C for 1.5 min, extension at 72 °C for 2 min, and a final extension step at 72 °C for 5 min. The PCR products (5 μL) were analysed by 1% agarose gel electrophoresis at 100 V for 30 min. The gels were stained with 0.4 μg/mL of ethidium bromide (Sigma, USA), and 100 bp DNA ladder (Promega, USA) was used as a molecular marker. The gels were examined under UV light using the AlphaImager HP System Versatile Gel Imaging (USA).

Isolates that were screened for E. coli based on phenotypic method and the presence of uidA gene were further molecularly confirmed through the matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) assay. The MALDI-TOF mass spectrometry (MS) device (Bruker Microflex LT, Germany) and Flex Control 3.0 software were used for identification by the mass spectrometer. In brief, each isolate was subcultured onto a nutrient agar plate. Then, using a sterile colony-transfer device, a single purified colony was streaked as a thin film directly onto a sample position on a cleaned 96 micro scout plate (MSP). Then the sample was spotted with 1 μL 70% aqueous formic acid and allowed to dry at room temperature (20–25 °C). Following drying, 1 μL of HCCA (α-cyano-4-hydroxycinnamic acid, Bruker) matrix solution was mixed and dried completely at room temperature. Next, the MALDI 96 MSP was loaded onto the MALDI-TOF MS device. Mass spectra were collected using the Bruker FlexControl 3.0 software (Bruker Daltonics) with a mass range of 2–20 kDa in the linear mode (laser frequency, 20 Hz; ion source 1 voltage, 20 kV). The experiments were done with 1,000 shots per spot. The m/z values of peaks were measured using the Bruker Biotyper 2.0 software (Bruker Daltonics), which provided smoothing, normalization, baseline subtraction, and peak picking. The resulting peak pattern was matched against reference patterns in the MALDI Biotyper reference library (Christner et al. 2014).

Double Disk Diffusion Synergy Test (DDST) was performed to determine the ESBL phenotypes of E. coli isolates. Here, bacterial culture was prepared with turbidity following the 0.5 McFarland standard, and then, the test inoculum was cultured on Mueller–Hinton agar (MHA) plates. An augmenting disk (20 μg amoxicillin + 10 μg clavulanic acid) along with three other antibiotic disks of ceftriaxone (30 μg), ceftazidime (30 μg), and cefotaxime (30 μg) were placed on the surface of the MHA plates, ensuring that each disk was 20 mm apart from the augmenting disk (centre to centre). Incubation was done at 37 °C for 24 h and then the zone of inhibition was measured in diameter. A ≥5 mm increase in zone diameter around the disks combined with clavulanic acid as well as the same disks alone confirmed the presence of ESBL phenotypes, as described by the Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI 2012).

The antibiotic resistance pattern of the isolates was determined by the standard Kirby–Bauer disk diffusion assay using MHA medium (Oxoid Limited, England), according to the CLSI guidelines (2020) (Bauer et al. 1966). We used 16 antibiotic disks from eight different groups, namely, Ampicillin (AM), Amoxycillin + Clavunic acid (AMC), Azithromycin (AZM), Amikacin (AK), Gentamycin (GN), Chloramphenicol (C), Cefixime (CFM), Cefuroxime (CXM), Ceftriaxone (CRO), Ceftazidime (CAZ), Cefotaxime (CTX), Ciprofloxacin (CIP), Levofloxacin (LEV), Nalidixic acid (NA), Meropenem (MEM), and Tetracycline (TE). These antibiotics were chosen for the study based on their frequent use and production in Bangladesh.

MDR pattern of the bacteria was determined by observing the resistance phenotypes against the used antibiotics. Multiple antibiotic resistance (MAR) indices were calculated for the isolates using the formula as MAR Index = a/b, where ‘a’ denotes the number of antimicrobials to which the bacteria was resistant and ‘b’ denotes the total number of antibiotics assayed (Kurdi Al-Dulaimi et al. 2019; Abayneh et al. 2023).

Beta-lactam group antibiotic cefixime was selected for determining the minimal inhibitory concentration (MIC) values for the study E. coli isolates. The agar dilution method was followed in accordance with the guidelines established by the CLSI (CLSI 2012). Quality control assays were carried out using the E. coli strain ATCC 25922. In summary, the isolates were subcultured at least 24 h prior to MIC testing. MHA (Oxoid) was prepared according to the manufacturer's specifications. Cefixime was obtained as a laboratory-grade standard powder and reconstituted according to the manufacturer's recommendations. Cefixime solution was freshly prepared and stored at −20 °C.

PCR was used to detect the presence of ESBL resistance genes bla_TEM, bla_SHV, and bla_CTX-M-13 using three distinct primer combinations. The PCR mixture volume and concentration were consistent with the previously established protocols, differing only in the primer sets and thermal cycling conditions employed for the amplification of each resistance gene (Ejaz et al. 2021; Son et al. 2021; Islam et al. 2023). The PCR for the bla_TEM gene involved an initial step at 94 °C for 2 min, followed by 30 cycles of 94 °C for 1 min, 60.5 °C for 1 min, 72 °C for 1 min, and a final extension at 72 °C for 10 min. For the bla_CTX-M-13 gene, PCR commenced at 95 °C for 3 min, followed by 30 cycles of 95 °C for 1 min, 55 °C for 1 min, 72 °C for 1 min, and a final extension at 72 °C for 5 min. In the case of the bla_SHV gene, PCR was initiated at 94 °C for 5 min, followed by 30 cycles of 94 °C for 1 min, 60 °C for 1 min, 72 °C for 1 min, and a final extension at 72 °C for 7 min. All the PCR reactions were performed in 25 μL reaction mixture consisting of 10.5 μL nuclease-free water, 12.5 μL G2 green master mix, 0.5 μL forward primer, 0.5 μL reverse primer, and 1 μL of DNA template. The PCR products were observed by 1% agarose gel electrophoresis, as described previously.

Multiplex polymerase chain reaction (mPCR) was performed for detecting the virulence genes eae, ipaH, eagg, and Lt of E. coli using four sets of primers (Mbanga et al. 2023). The thermal cycler utilized for this process was the 2720 Thermal Cycler from Applied Biosystems, USA. The procedure initiated at 94 °C for 10 min as the denaturation step. Subsequently, 35 amplification cycles were performed, comprising 40 s of denaturation at 94 °C, 30 s of annealing at 55 °C, and 50 s of extension at 72 °C. The final extension step was conducted at 72 °C for 7 min. The anticipated amplified products for the virulence genes were assessed through 2% agarose gel electrophoresis. All the primers used in the study for assessing ESBL resistance and virulence genes are listed in Table 1.

All the data were analysed statistically using the GraphPad prism software (version 8.0.1). Association between the categorical data were examined using Pearson's Chi-square test and Fisher's exact test with a significance level of 0.05.

The temperature of the water samples was found to be in the range of 22.3–30.8 °C, whereas the pH levels ranged between 6.86 and 8.32. Bacterial count of the samples ranged from 40 × 10⁴cfu/mL to too numerous to count. The total Gram-negative bacterial count was found in the range of too few to count to 81 × 10⁴cfu/mL. In case of the total ESBL-producing bacterial count, for the control samples no colony was found, whereas for the rest of the samples, the count was in the range from 5 × 10⁴ to 48 × 10⁴ cfu/mL. The highest number of ESBL-producing bacterial isolates were found in the textile industries surrounding the sources of the water samples. All the sample sources had significant statistical association with the prevalence of bacterial isolates and Gram-negative bacteria except two, which were collected near agriculture fields as per the Chi-square test (Table 2).

Initially, we selected 96 isolates based on their characteristic colony morphology as E. coli, from which 68 isolates were further confirmed based on the phenotypic metallic sheen on EMB agar and presence of uidA gene in the PCR. The isolates were further confirmed to species level by MALDI-TOF, which provided a score list in the range of 2.13–2.39. Among these 68 isolates, seven isolates were from the control water samples.

In Bangladesh, many areas like Savar lack adequate and proper wastewater treatment facilities. As a result, untreated or poorly treated wastewater containing antibiotics, antibiotic residues, and antibiotic-resistant bacteria is often discharged directly into rivers, lakes, and other water bodies, further exacerbating the problem of MDR in the aquatic environment. Monitoring multidrug-resistant E. coli in water sources affected by pollution activities is essential for safeguarding public health, understanding antibiotic resistance dynamics, and promoting sustainable environmental management practices in Bangladesh. From these viewpoints, we aimed to assess the drug resistance profile of ESBL-producing E. coli in water bodies near different industrial, agricultural, and farming practices in Bangladesh.

The physiochemical parameters such as temperature and pH of the study water samples compiled the limit as per WHO and the Department of Energy (DoE) standard for Bangladesh (WHO 1993; DoE 2001). This finding coincides with some other studies reported in Bangladesh (Ahmed et al. 2010; Hasan et al. 2019; Khan et al. 2020). In terms of heterotrophic bacteria, all the samples had higher counts. This higher load of bacteria could be influenced by the plethora of nutrients or microbial pollution released from the surrounding activities in the respective sources studied in the present investigation. The higher heterotrophic bacterial count sometimes could be an early pollution indicator (Bisi-Johnson et al. 2023). Besides, higher heterotrophic bacterial number may deteriorate the quality of water as they decompose the effluents from other sources and deplete oxygen rapidly (Garnier et al. 1992). Similarly, all the water samples had higher Gram-negative bacterial count, which implies that the water is unsuitable for any domestic and recreational purposes. Further, we have found higher number of ESBL bacteria in wastewater from pharmaceutical companies, poultry farms, and textile industries, indicating these waters contain more resistant bacteria than other sources. The results for this bacterial count concur with some other studies reported in Bangladesh (Saha et al. 2009; Real et al. 2017). For the total ESBL bacteria count, the samples of this study had fewer numbers than an earlier study on environmental water of Dhaka, Bangladesh (Haque et al. 2014).

The Chi-square analysis in Table 2 reveals inconsistency in bacteriological water quality across different sample sources. Significant p-values (<0.001) indicate significant differences in bacterial counts between sampling sites within the same source, which suggest the variation in pollution levels. For example, the considerable variances reported in pharmaceutical water (PhW) and poultry water (PW) indicate heterogeneity in bacterial pollution rather than other sources. By contrast, non-significant p-values (0.199) indicate consistent bacterial counts, implying more homogeneous water quality in the field water samples. Overall, the Chi-square assay demonstrates that some water sources had greater variation in bacteriological quality, presumably pointing to different levels of pollution or other influencing factors between sampling sites.

In this study, 96% (n = 61) of the total E. coli isolates were multiple drug resistant with the MAR values ranging between 0.3 and 0.9, whereas none of the E. coli from the control samples had a multiple drug resistance pattern. This indicates the impact of pollution activities on the development of drug resistance in the surrounding microbiota. Also, the MAR values for the MDR E. coli isolates greater than 0.2 stipulate the possibility of high-risk contamination of antibiotics in the sampling sources (Ayandele et al. 2020). For the individual sampling sources, MDR percentage was 100% for the pharmaceutical industries and textile industries, whereas it was 94% for poultry farms and agricultural lands. In our earlier investigation, we discovered 100% MDR among E. coli isolates from water surrounding pharmaceutical industries (Mou et al. 2023). Due to lack of proper treatment prior to disposal, wastewater from pharmaceutical industries contribute in the development of higher level of drug resistance among the bacterial isolates. Although antibiotics are not used in the textile industries, higher percentage of MDR E. coli isolates from these sources could be associated for two reasons. First, the textile industries selected as sampling sites in our study were adjacent to many medical facilities in Savar and the waterbodies are mostly interconnected. Wastewater treatment is also insufficient in the hospitals of Bangladesh, which might in turn disseminate the drug resistance into the surrounding environment. Second, in Bangladesh, the textile industries use different heavy metals and dyes, which could result in the resistance of these toxic substances to the surrounding bacterial biota. Co-resistance to antibiotics could be developed from the continuous exposure of the bacterial isolate's selection pressure from these pollutants (Ahmed et al. 2020).

The most frequent MDR pattern of the isolates was resistance to Ampicillin–Azithromycin–Nalidixic acid, which is similar to the findings of our previous study on E. coli isolates from pharmaceutical industries surrounding surface water (Mou et al. 2023). A higher percentage of ampicillin resistance is linked with their long-term overuse in Bangladesh, especially in the poultry sector (Mahmud et al. 2020; Ibrahim et al. 2023). This finding is also relevant to some other studies reported worldwide (Edge & Hill 2005; Maal-Bared et al. 2013). Azithromycin is also a frequently used macrolide group of antibiotics in Bangladesh, particularly for treating invasive E. coli infection in humans and in animal farms (Ibrahim et al. 2023). From these animal farms, the antibiotic resistance could be disseminated to the surrounding environments including the agricultural fields. The extensive use could be attributed to greater level of azithromycin resistance of E. coli in this study. A previous study on cattle and their environment in Bangladesh reported 100% resistance to Azithromycin of E. coli (Al Amin et al. 2020). Despite the fact that the Directorate General of Drug Administration in Bangladesh banned the use of azithromycin in animal enterprises in 2022, antibiotics are still available to buy and use without any regulations (DGDA 2022). Nalidixic acid is not very commonly used in poultry or veterinary feeds. This antibiotic is misused therefore by its random distribution and over-the-counter sales without prescription (Ahmed et al. 2019; Ibrahim et al. 2023). Here in our study, the prevalent MDR pattern was also higher for the isolates from the water samples near agricultural fields followed by those near poultry farms. This is certainly linked to persistent irrational use of antibiotics in the poultry and animal farm sectors of Bangladesh. The lowest percentage of resistance was found for meropenem of the E. coli isolates in our study. Meropenem is an infrequently used antimicrobial agent in Bangladesh, therefore this lower percentage of resistance is relatedly expected for the E. coli isolates (Hossain et al. 2020; IQVIA 2021). This finding is supported by other studies on E. coli isolates of Bangladesh (Rashid et al. 2015; Mahmud et al. 2020).

In the current study, the majority of the isolates were bla_TEM gene positive followed by bla_CTX-M-13. This is perfectly consistent with our prior study reported for E. coli isolates from water near pharmaceutical companies of the Savar area (Mou et al. 2023). Through another study, we found the predominance of bla_SHV gene among the enterobacteria from surface water sources of Bangladesh (Rahman et al. 2004). However, dominance of bla_TEM and bla_CTX-M-13 coincide with the findings of some other studies in Bangladesh (Qadri et al. 2005; Mahmud et al. 2020). E. coli isolates harbouring one, two, or a combination of the three studied ESBL genes were observed with MDR phenotypes non-susceptible to three or more than three groups of antibacterial agents used in this study. Table 4 compares the findings of genotypic and phenotypic assays, with higher resistance rate in the phenotypic assay rather than the genotypic assay. This suggests that while the specific resistance genes (bla_TEM, bla_CTX, and bla_SHV) may not always be present in the bacterial isolates, they could be phenotypically resistant to those groups of antibiotics, which might be due to other mechanisms or undetected resistance genes. The seven bla_TEM-positive strains mentioned in the table are all resistant according to the phenotypic test. This means that when the bla_TEM gene is present in these strains (detected genotypically), they also showed resistance in phenotypic assay. This indicates a strong correlation between the presence of the bla_TEM gene and the phenotypic expression of resistance. The p-value compares the distribution of resistance and susceptibility in the phenotypic assay with the findings of genotypic assay positive for any of the three genes (bla_TEM, bla_CTX, bla_SHV) and the combinations of two/three genes. The p-value indicates that the inconsistency between the genotypic and phenotypic data is not due to random chance. Specifically, the p-values show that phenotypic resistance is significantly higher than what genotypic testing alone would determine. This implies the importance of both phenotypic and genotypic methods for a more accurate observation of antibiotic resistance of the bacterial isolates.

Phenotypic and genotypic detection of the ESBL resistance among the bacterial isolates was assessed statistically with regard to their sources (Table 5). It indicated a strong correlation between phenotypic and genotypic resistance in pharmaceutical, textile, and poultry samples, suggesting that these environments are significant reservoirs of resistant bacteria. Agricultural fields also show a high correlation, indicating potential risks associated with antibiotic resistance in these environments. Control samples, by contrast, show minimal resistance, underscoring the impact of anthropogenic activities on resistance patterns. Further research is needed to explore the mechanisms driving resistance in these environments and to develop strategies to mitigate the spread of ESBL genes.

Although the E. coli isolates from the control water sources were not multidrug resistant, five isolates were found to be non-susceptible to one or two antibiotics tested in our study. Resistance in these isolates could be due to intrinsic resistance or horizontal gene transfer, which could be better understood through further molecular investigation of the isolates. As we explored the presence of only three ESBL resistance genes, in future PCR of other ESBL and antibiotic resistance genes could shed light on the exact scenario. Moreover, sometimes the water sources could be the natural reservoir of antibiotic resistance, which could be the reason for non-susceptible phenotypes in control water E. coli isolates. Out of 61 E coli isolates in this study, some were observed with phenotypic sensitivity to the tested antibiotics, while they contained the ESBL resistance gene. This could be due to silencing or the mutation of the resistant genes. This type of discrepancy between phenotypic and genotypic resistance has also been reported in some previous studies (Davis et al. 2011; Rasheed et al. 2023). Similarly, many of the E. coli isolates were not found to harbour the resistant gene, but showed phenotypic resistance in disk diffusion assay. As mentioned previously, in our study we did PCR for only a few specific resistance genes, so there is a possibility that they may contain other variants of the resistance genes which were not included in our study. Additionally, the silence of some resistance genes could also potentially lead to the observed discrepancies between the phenotypic and genotypic resistance patterns (Davis et al. 2011). Out of 61 E. coli isolates near polluted sources, seven were virulent ones, with the prevalence of EAEC followed by ETEC. A similar pattern was observed in other reported investigations of Bangladesh (Lothigius et al. 2008; Mahmud et al. 2021). Previously, it has also been demonstrated that virulent types of E. coli can be found in both drinking water and environmental water in Bangladesh (Lothigius et al. 2008). Furthermore, it remains viable even after prolonged incubation in water, indicating that water could serve as a significant mode of transmission.

This study found a higher prevalence of resistance to several antibiotics, as well as ESBL resistance genes, in environmental E. coli isolates, highlighting the grave AMR situation in Bangladesh. Irrational antibiotic usage without a prescription is frequent here, but facilities for AMR surveillance and detecting MDR and ESBL genes are insufficient. In densely populated countries like Bangladesh, inadequate waste treatment and their disposal systems also significantly contribute to the emergence and proliferation of antibiotic resistance. The water bodies could be a significant reservoir for the transmission and propagation of MDR, ESBL-producing, and pathogenic variants of E. coli, posing substantial public and environmental health risks. Consequently, based on the outcomes of this research, we strongly recommend policymakers prioritize the implementation of robust control strategies, particularly focusing on monitoring antibiotic usage and proper disposal practices to mitigate the spread of drug resistance in the environment.

This study has been supported by grants from the University Grant Commission (UGC), Bangladesh, and Jahangirnagar University, Bangladesh.

Conceptualization and designing of experiments: T.J.M.; Investigation: S.H.S., N.A.N.; Data analysis: S.H.S., N.S., T.J.M.; Manuscript preparation: T.J.M., S.H.S., F.I.; Funding acquisition: T.J.M., F.I., A.K.P.; Validation: A.K.P., S.K.D.; Supervision: T.J.M., A.K.P.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.