1 Introduction

The worldwide emergence and spread of antimicrobial resistance (AMR) as a “silent pandemic” may result in up to 10 million AMR-related mortality by 2050 (Bengtsson-Palme et al., 2018; WHO, 20172019). Often the environment acts as a reservoir and “mixing vessel” of AMR pathogens that harbor AMR genes (ARGs) in their plasmids, integrons, transposons, or insertion sequences which can be transferred to other pathogenic and commensal microbes via horizontal gene transfer (Raimondi et al., 2019; Ramos et al., 2020). Among the environmental components, the wastewater originating from various sources like hospitals, livestock and poultry farms, and slaughterhouses is a major source of antibiotic-resistant bacteria and resistant genes. These can enter natural water sources like rivers, lakes, and groundwater through the discharge of treated or untreated wastewater (Adelowo et al., 2018). This creates high selection pressure zones within the recipient aquatic environment which helps in further dissemination of AMR organisms with novel resistant mechanisms through recreational water activities or consumption of contaminated water or food (Amos et al., 2014; Alouache et al., 2014; Chukwu et al., 2023).

Escherichia coli is a common inhabitant of the gastrointestinal tract of humans and animals, and it has been used as a sentinel organism for antimicrobial surveillance studies (Nyirabahizi et al., 2020). Antibiotic resistance (AR) in E. coli to β-lactams (viz. penicillin, cephalosporins, carbapenems, etc.) is mainly mediated by the production of β-lactamase enzymes (Hussain et al., 2021; Sta Ana et al., 2021). Among the β-lactamases, the extended-spectrum beta-lactamases (ESBLs) are of special importance which confer resistance to all β-lactams except carbapenems and cephamycin and is inhibited by β-lactamase inhibitors like clavulanic acid, sulbactam, or tazobactam (Paterson & Bonomo, 2005). The worldwide dissemination of ESBL-producing E. coli in clinical infections, foods of livestock and plant origin, aquatic and soil ecosystems, and wastewater from industry, slaughterhouses, hospitals, livestock farms, and sewage treatment plants is a matter of serious concern for public health and environment ecosystem (Conte et al., 2017; Girijan & Pillai, 2023; Gregova et al., 2021; Korzeniewska et al., 2013; Marathe et al., 2019). Many clinically significant ESBL-producing E. coli harbor plasmid-mediated blaTEM, blaSHV, blaCTX-M, and blaOXA gene variants (Shaikh et al., 2015; Liakopoulos et al., 2016; Hooban et al., 2020). Infections caused by multi-drug resistant (MDR) ESBL-producing E. coli is challenging to treat and their occurrences in the environment possess a significant public health risk (Kasanga et al., 2023).

In the context of the “One Health” approach, the detection and characterization of the ESBL-producing bacteria at the human-animal-environment interface, taking E. coli as the indicator organism, can add substantial knowledge to safeguard public health. This will also preserve the effectiveness and use of antibiotics. In the existing literature, there is a lack of information related to the detection and characterization of ESBL-producing bacteria at the human-animal-environment interface, particularly from developing countries like India. Hence, the present study was undertaken to detect the occurrence of ESBL-producing E. coli in wastewater from hospitals and livestock farms as well as the natural water sources (viz. rivers, canals, reservoirs, etc.) from five districts of the eastern parts of Uttar Pradesh, India, based on their proximity and constant interaction. The recovered E. coli isolates were further tested by employing molecular tools for the detection of ESBL genes, phenotypic co-occurrence of ESBL-production and multi-drug resistance (MDR) traits, and phylogenetic typing. This study aims to provide an insight into the perpetuation of ESBL-producing E. coli in environmental sources in parts of Eastern Uttar Pradesh, India, which will add information on AMR surveillance on “One Health” perspective.

2 Materials and Methods

2.1 Study Area and Collection of Samples

The study was carried out in the Department of Veterinary Public Health and Epidemiology at the Faculty of Veterinary and Animal Sciences, Rajiv Gandhi South Campus, Banaras Hindu University, Mirzapur, Uttar Pradesh, India, from January 2022 to March 2023. A total of 134 sites comprising 58 livestock farms, 43 hospitals, and 33 natural water sources were included for sample collection by simple random sampling method (95% confidence interval, ± 5% precision applied to determine the number of sampling sites with respect to places and sources) from five districts (Prayagraj (25.43°N latitude and 81.84°E longitude), Mirzapur (25.12°N latitude and 82.56°E longitude), Varanasi (25.31°N latitude and 82.97°E longitude), Sonbhadra (24.53°N latitude and 83.03°E longitude), and Jaunpur (25.75°N latitude and 82.68°E longitude)) of eastern Uttar Pradesh, India, which are located within a 100-km radius around the place of work (Mirzapur) (Fig. 1, Supplementary Table 1). In each sampling event, approximately 500 mL of pooled samples (livestock farm wastewater (LFWW), hospital wastewater (HWW) as well as water samples from natural water sources (NWS)) was collected from three different spots of each site by grab sampling method. The water sample was collected in sterile wide-mouthed glass bottles preferably in the morning hours (Baird et al., 2017), transported to the laboratory in a cold chain (4 ± 1℃), and processed within 6 h.

Fig. 1
figure 1

Map revealing places and sources of sample collection from five districts (1, Prayagraj; 2, Mirzapur; 3, Varanasi; 4, Sonbhadra; 5, Jaunpur) of the eastern part of Uttar Pradesh, India. Livestock farm wastewater (LFWW) sources are represented by brown dots, hospital wastewater (HWW) sources as red dots, and natural water sources (NWS) as blue dots, respectively

2.2 Isolation of Cefotaxime-Resistant E. coli

The water sample was thoroughly mixed by inverting the bottle several times and 1 mL of the sample was enriched in 9 mL of McConkey broth and incubated overnight at 37℃. After incubation, a loopful of the inoculum was streaked on Eosin Methylene Blue agar plates supplemented with 4 µg/mL of cefotaxime and incubated overnight at 37℃. The colonies exhibiting a greenish metallic sheen with dark purple background were presumptively identified as Escherichiae. The phenotypic appearance of colonies may not reflect the genetic variability among different strains in the same sample, for which additional testing like antibiotic resistance profiling, serogrouping, plasmid profiling, or DNA fingerprinting is required. Also, it is difficult to ascertain the number of colonies to be picked from each plate without knowledge of the actual number of strains on the plate (Singer et al., 2000). Earlier study on the isolation of E. coli from cellulitis lesions in broilers processed three colonies from each McConkey plates with an assumption that there would be three or fewer strains per lesions and equal growth of different strains on the plates (Singer et al., 2000). Hence, at least three colonies were randomly selected from different streaking lines of the plates and characterized by biochemical tests and molecular tools for the identification of Escherichia coli. After molecular confirmation, these were further processed to determine the ESBL production by phenotypic and genotypic methods, antimicrobial resistant profile, and phylogenetic grouping.

2.3 Molecular Confirmation of E. coli by PCR Assay

The cefotaxime-resistant E. coli isolates were confirmed by polymerase chain reaction (PCR) assay by amplifying the housekeeping uidA gene as described previously (Alsanjary & Sheet, 2022; McDaniels et al., 1996). In brief, the genomic DNA was extracted by boiling lysis method (Mahmud et al., 2020). The PCR assay was carried out in a 25-µL reaction mixture comprising of 12.5 µl 2 × master mix (Promega Corporations, USA), 1 µL (10 pmol/µL) each of forward (5′-CCAAAAGCCAGACAGAGT-3′) and reverse (5′-GCACAGCACATCCCCAAAGAG-3′) primers (synthesized at Eurofins Genomics Pvt. Ltd., Bengaluru), 7.5 µL NFW (Promega Corporations, USA), and 3 µl DNA template. The PCR products were electrophoresed on 1.5% agarose gel (Genei, Bangalore) with a 100-bp DNA ladder (BR Biochem, India) and visualized in a gel documentation system (Biorad, USA).

2.4 Phenotypic Identification of ESBL Production by DDST, CDM, and MIC

All the PCR-confirmed cefotaxime-resistant E. coli isolates were screened for phenotypic production of ESBL enzymes by double disk synergy test (DDST) and combination disk method (CDM) using Mueller Hinton Agar (MHA) and the results were interpreted as per Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI, 2018). For DDST, ceftazidime (CAZ 30 µg) and/or cefotaxime (CTX 30 µg) disks and amoxicillin/clavulanate (20 µg/10 µg) disks (Himedia, India) were placed at a distance of 30 mm on sterile MHA plates swabbed with the standard inoculum (0.5 McFarland) of the test isolates. The plates were incubated overnight at 37℃. Extension of the zone of inhibition (ZOI) towards amoxicillin/clavulanate disk was measured to interpret ESBL producer (Drieux et al., 2008; Sageerabanoo et al., 2015). For CDM, ceftazidime (CAZ 30 µg) and/or cefotaxime (CTX 30 µg) disks along with ceftazidime-clavulanate (30 µg/10 µg) and/or cefotaxime-clavulanate (30 µg/10 µg) disk (Himedia, India) respectively were placed at a distance of 30 mm on sterile MHA plates swabbed with the standard inoculum (0.5 McFarland) of the test isolates. The plates were incubated overnight at 37℃. An increase in the ZOI ≥ 5 mm around the disk with the antibiotic-clavulanic acid combination compared to the antibiotic disk alone was considered positive for ESBL production (CLSI, 2018).

The minimum inhibitory concentration (MIC) of antibiotics (ceftazidime and cefotaxime alone and their combination with clavulanic acid) was determined by microbroth dilution assay in cation-adjusted Mueller Hinton broth (CAMHB) using ceftazidime (0.25–512 µg/mL), ceftazidime plus clavulanic acid (0.25/4–512/4 µg/mL), and cefotaxime (0.25–512 µg/mL), cefotaxime plus clavulanic acid (0.25/4–512/4 µg/mL) in 96-well microtiter plates (Corning, USA) (CLSI, 2018). The plates were incubated for 18–24 h at 37℃. A ≥ 3 twofold concentration decrease in a MIC for either antibiotic tested in combination with clavulanic acid vs the MIC of the antibiotics when tested alone was considered phenotypically positive for ESBL production. As quality control strains, Klebsiella pneumoniae ATCC 700603 and E. coli ATCC 25922 were used.

2.5 Molecular Detection of Plasmid-Mediated ESBL Genes Among the Phenotypically ESBL-Positive Isolates

Plasmid DNA was extracted from all the phenotypically positive ESBL-producing E. coli isolates (overnight grown fresh cultures) employing the alkaline lysis method (Sambrook & Russell, 2001) and subjected to molecular detection of plasmid-mediated ESBL-producing genes by multiplex-PCR (mPCR) assay as described previously (Dallenne et al., 2010) using primers (synthesized at Eurofins Genomics Pvt. Ltd., Bengaluru, India) presented in Supplementary Table 2. The mPCR assay was carried out in two sets—the first one amplified the target regions of blaTEM, blaSHV, and blaOXA-1-like genes, and the second one amplified that of blaCTX-M gr-1, blaCTX-M gr-2, blaCTX-M gr-8, blaCTX-M gr-9, and blaCTX-M gr-25 genes in separate 25-µL reaction mixtures comprising of 12.5 µl 2 × master mix (Promega Corporations, USA), 1 pmol/µL of respective primers, 2 µl template DNA, and NFW to adjust the volume. The PCR product was electrophoresed and visualized as described earlier.

2.6 Antibiotic Sensitivity Test and MAR Index

All the genotypically positive ESBL-producing isolates were tested for their sensitivity to other antibiotics by the Kirby-Bauer disk diffusion method. Briefly, the sterile MHA plates were swabbed with the standard inoculum (0.5 McFarland) of the test isolates. The antibiotics used in this test belonged to nine different classes, namely penicillin: ampicillin (AMP 30 µg), piperacillin (PI 100 µg), and piperacillin-tazobactam (PIT 100 µg/10 µg); cephalosporins: cefepime (CPM 30 µg) and cefoxitin (CX 30 µg); aminoglycosides: gentamicin (GEN 10 µg) and kanamycin (K 30 µg); macrolides: azithromycin (AZM 15 µg); tetracyclines: doxycycline (DO 30 µg); folate pathway antagonists: co-trimoxazole (COT 1.25/23.75 µg); quinolones and fluoroquinolones: ciprofloxacin (CIP 5 µg) and nalidixic acid (NA 30 µg); phenicols: chloramphenicol (C 30 µg); and nitrofurans: nitrofurantoin (NIT 300 µg). The results were interpreted according to CLSI breakpoints (CLSI, 2018). Isolates that displayed resistance to three or more antibiotic classes were considered multi-drug resistant (MDR) (Magiorakos et al., 2012). The multiple antibiotic resistance (MAR) index was determined as the ratio of the number of antibiotics to which the isolate was resistant to the total number of antibiotics used in the test (Krumperman, 1983).

2.7 Phylogenetic Typing of ESBL-Producing E. coli Isolates

All the genotypically positive ESBL-producing E. coli isolates were subjected to phylogenetic typing to determine the distribution of phylogroups (A, B1, B2, C, D, E, and F) using PCR assay as described previously (Clermont et al., 2013).

2.8 Statistical Analysis

The difference in phenotypic ESBL production among cefotaxime-resistant E. coli isolates recovered from various sources and places were compared by the Pearson chi-square test. The presence of ESBL genes among phenotypically ESBL-producing E. coli isolates concerning their source and place of sample collection was compared by Pearson chi-square (when n > 3) and Fisher’s exact test (when n < 3) using IBM SPSS version 23. The level of significance was determined at a 95% confidence interval, where p < 0.05 was considered significantly different and p > 0.05 was non-significantly different.

3 Results

3.1 Phenotypic Identification of ESBL-Producing E. coli

In this study, a total of 310 cefotaxime-resistant E. coli isolates were recovered from various environmental sources. Their distribution across places and sources is presented in Table 1 and Fig. 2. All the isolates were biochemically and molecularly confirmed as E. coli by PCR assay targeting species-specific uidA gene which yielded a 623-bp product upon electrophoresis.

Table 1 Distribution of phenotypic ESBL-producing E. coli isolates recovered from various environmental sources collected from the eastern part of Uttar Pradesh, India
Fig. 2
figure 2

Distribution of phenotypic ESBL-producing E. coli isolates recovered from various environmental sources (LFWW, livestock farm wastewater; NWS, natural water sources; HWW, hospital wastewater; n, number of samples from respective sources)

Of the cefotaxime-resistant E. coli isolates, 91.29% (283/310) were observed to be phenotypically positive for ESBL production based on the DDST and CDM results (Table 1, Supplementary Fig. 1 and 2). A highly significant difference (p < 0.01) was observed in the occurrence of phenotypic ESBL production among the cefotaxime-resistant E. coli isolates recovered from different places (districts) and sources (LFWW, HWS, and NWS) of the Eastern Uttar Pradesh (Table 2 and 3). After Bonferroni correction, a significant (p < 0.01) number of cefotaxime-resistant E. coli isolates exhibiting phenotypic ESBL production and those not exhibiting ESBL production were obtained from LFWW and NWS sources than HWW sources (p > 0.05). Further, phenotypic ESBL expression among cefotaxime-resistant E. coli isolates recovered from Jaunpur and Sonbhadra districts differed significantly (p < 0.05), although a similar significant difference was not evident (p > 0.05) among cefotaxime-resistant E. coli isolates recovered from Prayagraj, Mirzapur, and Varanasi.

Table 2 Statistical significance of phenotypic ESBL production among cefotaxime-resistant E. coli isolates according to sources of sample
Table 3 Statistical significance of phenotypic ESBL production among cefotaxime-resistant E. coli isolates according to the place of sample collection

The MIC of cefotaxime and ceftazidime for the phenotypically ESBL-positive E. coli isolates was found to be in the range of 64–512 µg/mL and 16–512 µg/mL, respectively (Table 1). A ≥ 3 twofold decrease in the MIC of antibiotics was observed when tested in combination with clavulanic acid (at a fixed concentration of 4 µg/mL). MIC values above 256 µg/mL for cefotaxime and ceftazidime were observed in 83.45% and 53.11% of E. coli isolates respectively. The highest MIC value for the antibiotics (512 µg/mL) was observed in isolates primarily recovered from LFWW sources.

3.2 Molecular Detection of ESBL-Producing Genes

One or more plasmid-mediated ESBL genes among phenotypically positive ESBL-producing E. coli isolates (n = 283) were detected in 38.51% (109/283) of the isolates by multiplex PCR assay (Table 4). Among the isolates (n = 109), the blaCTX-M-gr-1 gene was detected predominantly (90.82%), followed by blaTEM (73.39%), blaOXA-1 like (11%), and blaSHV (2.75%) genes. Co-occurrence of blaCTX-M-gr-1 and blaTEM genes was observed in 51.37% (56/109) of isolates (Table 4). ESBL genes could not be detected in 61.48% (174/283) of phenotypically positive ESBL-producing E. coli isolates. Also, a significant difference (p < 0.05) was observed in the presence of blaCTX-M-gr-1, blaTEM, and blaOXA-1-like genes among phenotypically positive ESBL-producing E. coli isolates and their sources, except for blaSHV gene (p > 0.05) (Table 5).

Table 4 Distribution of ESBL genes among phenotypic ESBL-producing E. coli isolates recovered from environmental samples according to place and sources of collection
Table 5 Statistical significance of the presence of ESBL genes among phenotypically positive ESBL-producing E. coli isolates according to sources of sample

Phenotypically positive ESBL-producing E. coli isolates recovered from LFWW sources revealed the distribution of blaCTX-M-gr-1 gene to the tune of 11.11% (2/18), 10.86% (5/46), 21.42% (6/28), 34.48% (10/29), and 57.14% (12/21) from Mirzapur, Varanasi, Prayagraj, Jaunpur, and Sonbhadra, respectively (Table 4). Besides, blaCTX-M-gr-1 gene was predominantly detected in all the ESBL-producing isolates (19/19) recovered from HWW sources from Mirzapur, whereas its detection was between 26 and 36% in ESBL-producing isolates recovered from the same source (HWW) of four other places. While ESBL-producing isolates were recovered from NWS, the blaCTX-M-gr-1 gene could be detected up to 69.23% (9/13) in isolates recovered from Varanasi, followed by that of Prayagraj (50%), Jaunpur (33.33%), Sonbhadra (33.33%), and Mirzapur (23.07%), respectively (Table 4).

The presence of the blaTEM gene was predominantly observed in isolates recovered from LFWW sources of the Sonbhadra district (52.38%, 11/21) and HWW sources of the Varanasi district (48%, 12/25). The distribution of blaOXA-1-like genes was limited to the isolates recovered from HWW sources from Mirzapur (15.78%, 3/19) and Jaunpur (22.22%, 2/9), although HWW (15%, 3/20), NWS (25%, 3/12), and LFWW (3.5%, 1/28) sources also revealed blaOXA-1-like genes from Prayagraj district. Moreover, the blaSHV genes were detected in only three isolates recovered from LFWW sources of Sonbhadra (Table 4). Overall, a significant difference (p > 0.05) was not observed concerning the occurrence of ESBL genes among the phenotypically positive isolates of different sampling places (Table 6).

Table 6 Statistical significance of the presence of ESBL genes among phenotypically positive ESBL-producing E. coli isolates according to place of sample collection

3.3 Antibiotic Sensitivity Test of ESBL-Producing E. coli Isolates

Antibiotic sensitivity test of ESBL-producing E. coli isolates (n = 109) revealed an MDR trait in 91.41% (104/109) of isolates; i.e., these isolates were resistant to three or more than three classes of antibiotics (Table 7, Supplementary Table 3). Overall antibiotic susceptibility pattern revealed all isolates (100%) to be resistant to ampicillin, piperacillin, cefotaxime, and ceftazidime irrespective of their place and source of collection. Apart from the β-lactam antibiotics, a 100% resistance was observed towards azithromycin by all isolates recovered from Varanasi, Prayagraj, and Sonbhadra irrespective of their sources of collection. Among the isolates recovered from LFWW sources from various places under this study area, a lower resistance was observed towards doxycycline (0–25%, except those from Varanasi—60%), chloramphenicol (0–46.67%), and gentamicin (0–45.45%). Resistance to doxycycline was observed to be the least (0–33.33%) among the isolates recovered from HWW sources across all places that were sampled.

Table 7 Antibiotic-resistance (%) of genotypically positive ESBL-producing E. coli isolates (n = 109) recovered from various environmental sources of the eastern parts of Uttar Pradesh, India

Based on the resistance profile of the isolates, a varied distribution of resistant patterns was observed. Resistance patterns of AMP-PI-PIT-K-CPM-CX-AZM-CIP-NA-CTX-CAZ and AMP-PI-PIT-K-CPM-AZM-CIP-NA-CTX-CAZ were the most frequently observed in 7 (6.54%) and 5 (4.67%) of the isolates respectively. The MAR index of all the isolates (100%) was found to be higher than 0.2 (0.31 to 1). An isolate (JC8a), recovered from an LFWW source of Sonbhadra, revealed a MAR index of 1.0 (resistant to all the antibiotics tested). Similarly, two isolates, viz., JC4a and VH9c which were recovered from an LFWW sample from Jaunpur and an HWW sample from Varanasi, respectively, had a MAR index of 0.94 (resistance to 15 antibiotics) (Supplementary Table 3).

3.4 Phylogenetic Typing of the Isolates

Phylogenetic groups identified among the ESBL-producing E. coli isolates (n = 109) included phylogroup A (40.37%, 44/109), B1 (27.44%, 31/109), B2 (1.83%, 2/109), C (11.93%, 13/109), D (4.59%, 5/109), E (2.75%, 3/109), and F (0.92%, 1/109). However, the remaining 10 isolates could not be typed by the current method (Table 8).

Table 8 Distribution of phylogroups among genotypically positive ESBL-producing E. coli isolates (n = 109) recovered from various environmental sources

4 Discussion

This is the first report on the detection of MDR ESBL-producing E. coli in wastewater from livestock farms, hospitals, and the natural water sources of the surrounding areas of five selected districts of the Eastern parts of Uttar Pradesh, India. In this study, about 91.29% of the recovered cefotaxime-resistant E. coli isolates were observed to be phenotypically positive for ESBL production. The use of a selective medium and picking more than one colony might be a probable reason for the higher isolation rate in this study, which is supported by similar observations by other researchers (Diwan et al., 2012; Samanta et al., 2018). This high detection rate of cephalosporin-resistant E. coli from wastewater and water sources is a major threat to the environment (Korzeniewska et al., 2013). A similarly high prevalence of cefotaxime resistance (100%) was observed among ESBL-producing E. coli (24/38) recovered from wastewater from a University Health Centre facility in Nigeria (Adekanmbi et al., 2020). In India, wastewater treatment and management are not commonly practiced in many hospital settings (Diwan et al., 2012). The situation is even worse when it comes to animal husbandry practices. Also, β-lactams are the most frequently prescribed antibiotics in India (Van Boeckel et al., 2014). Most of the veterinary antibiotics (up to 90%) could not be assimilated in animals and excreted unchanged via urine or feces (Kemper, 2008), which might contaminate the nearby surface or groundwater. It has been reported that the use of third- and fourth-generation cephalosporins in livestock 12 months before the analysis may ensure at least four times more ESBL-positive E. coli isolates (Snow et al., 2012). Even the use of antibiotics other than β-lactams has been known to select ESBL phenotypes in E. coli isolates recovered from cattle farms (Schmid et al., 2013). All these probable factors in addition to antibiotic over-prescription, irrational use, over-the-counter availability, incomplete dose regimen, lack of sanitation (Kalasseril et al., 2020; Manyi-Loh et al., 2018), and overcrowding with a floating population at the pilgrimage sites (e.g. Prayagraj and Varanasi) of eastern Uttar Pradesh region might have induced a high selection pressure in the environment leading to higher detection of MDR bacteria.

The predominant ESBL genes detected among ESBL-producing E. coli isolates from LFWW sources were blaCTX-M-gr-1 (87.50%, 35/40) and blaTEM (72.50%, 29/40) across all places under the study area. Similar observations have been reported in an earlier study from Thailand, wherein 60.2% of ESBL-producing E. coli isolates that were recovered from dairy farms wastewater revealed the presence of blaCTX-M and blaTEM genes (Saekhow & Sriphannam, 2021). Besides, the blaCTX-M gene was found to be the major genetic determinant in ESBL-producing E. coli isolated from floor swabs and drinking water of a pig pen in West Bengal, India (Samanta et al., 2018). In India, most of the studies on the detection of ESBL-producing E. coli have focused on foods of animal origin, clinical samples, and fecal samples (Ghatak et al., 2013; Pruthvishree et al., 2017; Rawat et al., 2018), Without proper drainage and disposal facilities, the surrounding environment of farms may get repeated contamination with farm wastewater which may increase the concentration and survivability of ESBL-producing E. coli (Hartmann et al., 2012). Further dissemination from those areas is possible via rainwater run-offs to nearby rivers or natural water sources which may increase the chance of human acquisition via drinking, bathing, or other recreational purposes (Runcharoen et al., 2017; Xi et al., 2009).

The majority of the ESBL-producing E. coli isolates recovered from HWW samples across different places in this study were observed to harbor blaCTX-M-gr-1, blaTEM, and blaOXA-1-like genes. A similar outcome was observed earlier (Girijan & Pillai, 2023), where blaCTX-M was the most prevalent (32.8%) ESBL-encoding gene followed by blaTEM (15.6%) among the E. coli isolates from clinical environments. A diverse distribution of ESBL genes (blaCTX-M-15, blaTEM-1 and blaSHV-12, blaOXA-10) among E. coli isolates from HWW samples have been observed in various studies from India (Diwan et al., 2012; Marathe et al., 2019; Bardhan et al., 2020; Kalasseril et al., 2020). Several studies have shown the emergence of blaCTX-M coded ESBLs (mainly the blaCTX-M-15 which is placed under the blaCTX-M-gr-1) as the most significant ESBL-encoding genes among cephalosporin-resistant E. coli in recent years from clinical isolates and hospital settings worldwide because of the overuse of cephalosporins (Cantón et al., 2012; D’Andrea et al., 2013; Ibrahimagić et al., 2015). The TEM-β-lactamases were also observed to be one of the most clinically relevant to the β-lactamase family (Girijan & Pillai, 2023). It has been reported that the frequency of ESBL producer and number of β-lactamase genes are directly associated with the “hospital size,” wherein bigger hospitals were found to contribute a greater bacterial load to their surrounding environment in comparison to smaller ones (Girijan & Pillai, 2023; Lamba et al., 2017). This was also evident in our study that the presence of high-end tertiary care hospitals in Prayagraj, Varanasi, and Mirzapur contributed a higher frequency of ESBL-producing E. coli and ESBL genes in comparison to that of Sonbhadra and Jaunpur. The continuous exposure of β-lactam antibiotics may be responsible for creating selective pressure for blaTEM and blaCTX-M genes, as a substantial increase in the prescription of certain β-lactam antibiotics like ceftriaxone, meropenem, and ertapenem has been linked to increased ARB in hospital effluent (Diwan et al., 2010; Hsu et al., 2010).

Screening of NWS samples from various places under this study has revealed the presence of blaCTX-M-gr-1 and blaTEM as the major genetic determinants of ESBL genes among the recovered isolates. Similar findings were observed in a study on river water samples collected from five states (Goa, Bihar, Karnataka, Tamil Nadu, and Telangana) of India (Akiba et al., 2016); Yamuna River water collected from Delhi (Bajaj et al., 2015), and Ganga river water collected across the state of Uttar Pradesh, India (Chaturvedi et al., 2020). Our study also supports the observations on the worldwide occurrence of CTX-M and TEM as the dominant ESBL-gene family in E. coli isolates from surface water (Blaak et al., 2015), steam and well water (Caltagirone et al., 2017), river water (Chen et al., 2010; Kamruzzaman et al., 2013; Kim et al., 2008; Wambugu et al., 2018; Zarfel et al., 2017), and recreational water (Jørgensen et al., 2017). Although large water bodies may dilute the concentration of antibiotics and metabolites, repeated exposure even in sub-therapeutic concentration provides an ideal milieu for the transfer of resistance among the bacterial community (Watkinson et al., 2009).

The current study detected mostly the blaTEM and blaCTX-M genes among the isolates in comparison to blaOXA-1-like or blaSHV genes. Similar observations have been noted in earlier studies where TEM, OXA, and CTX-M genes were predominantly detected from isolates recovered from wastewater samples (Kutilova et al., 2021; Liedhegner et al., 2022; Nzima et al., 2020). It may be hypothesized that bla genes have exchanged and recombined in the aqueous environment giving rise to predominantly blaTEM-and blaCTX-M genotypes. The bla gene transfer theory may explain the possible difference between clinical bla gene types and those obtained from the environment (Chen et al., 2010). Wastewater from animal farms may contain ESBL-E. coli with completely different resistant genetic determinants, but reports are scarce in this regard for discussion. The TEM enzyme has been known to be an ESBL-producing enzyme for a long time, but the significance of CTX-M enzymes has been recently known (Paterson & Bonomo, 2005). In our study, negative PCR results of ESBL genes were observed in 61.48% of phenotypically ESBL-positive isolates. This probably indicated the presence of other non-detectable ESBL-producing genes (like GES, PER, VEB, SFO, TLA) in the same molecular class A or novel ESBL genes or substantial genetic variations. A similar outcome was observed in a study on the blaVIM carbapenemase gene among phenotypic carbapenem-resistant E. coli isolates recovered from calves in India (Murugan et al., 2019). Additionally, the β-lactamase gene selected and amplified in this case represents a selected cellular mechanism of drug resistance deployed by prokaryotes. Moreover, the involvement of other mechanisms in drug resistance cannot be ruled out. In the absence of sequence data, it was not possible to interpret whether a lower detection rate of ESBL genes was due to a problem in the confirmation method or the presence of another enzyme not detected by the current PCR-based method (Bell et al., 2007).

In this study, about 96.33% (105/109) of isolates exhibited an MDR pattern with a MAR index above 0.3. Similar findings were also obtained in a study from Ecuador where 98.3% of cefotaxime-resistant isolates from broiler farms were MDR (Vinueza-Burgos et al., 2019). The occurrence of MDR isolates in environmental sources is highly worrisome from a public health point of view. It has been reported that bacteria having a MAR index ≥ 0.2 originate from high-risk areas where several antibiotics are used (Afunwa et al., 2020; Sandhu, 2016). Most of the isolates exhibited high resistance to the quinolone class of antibiotics. Co-existence of cephalosporin and quinolone resistance phenotypes and genotypes among Enterobacteriaceae was previously reported (Galvin et al., 2010; Lavilla et al., 2008) as the plasmids carrying blaCTX-M and qnr genes occur as self-transmissible plasmids carrying transportable ARGs (Diwan et al., 2012). Isolates from HWW samples of Prayagraj, Varanasi, and Mirzapur were observed to be resistant to Kanamycin, which may be due to their higher indiscriminate use in those areas. It is noteworthy and a matter of concern that 42–93% of E. coli isolates recovered from HWW sources of Prayagraj, Varanasi, and Mirzapur was resistant to cefepime, a fourth-generation cephalosporin. One of the interesting findings is 96% (104/109) of the ESBL-producing E. coli isolates were co-resistant to azithromycin irrespective of the place or source of sample collection. It can be hypothesized that increased and indiscriminate use of azithromycin in the treatment of COVID-19 infection and other respiratory illnesses during the ongoing pandemic might have induced an extreme selection pressure on bacteria leading to their higher resistance to azithromycin. This is well supported by various studies wherein an increased use of azithromycin has led to bacterial resistance to this drug (Seabra et al., 2021; Silva Segundo et al., 2022; Sultana et al., 2020; Wu et al., 2020). Azithromycin has been increasingly and widely used in the management of COVID-19 (Ayerbe et al., 2022; Gyselinck et al., 2021). It was also observed that an increase in use of azithromycin (a macrolide) in COVID-19 patients had led to an increased resistance to erythromycin (another macrolide) in Staphylococcus aureus in Mexico (López-Jácome et al., 2022). Moreover, another study from Spain reported that 10 COVID-19 patients out of 48 (21%) admitted to ICU revealed azithromycin resistance strains of P. aeruginosa, E. faecium, H. influenza, and MRSA (Barrasa et al., 2020). Resistant determinants against aminoglycosides, tetracyclines, sulphonamides, and cephalosporins are often present on the same plasmid along with resistance to some heavy metals and toxic elements which might have aided in the co-occurrence of ESBL and MDR traits displayed by the isolates in our study. This will provide a suitable platform for the exchange of genetic material among the microbiota through horizontal gene transfer leading to the spread of MDR strains across the anthropogenic environment (Ibrahim et al., 2016; Chaturvedi et al., 2020). Comparison with other studies may not give the same resistant pattern, but studying resistance to different groups of antibiotics in a sentinel species like E. coli has indicated the presence of an antibiotic “resistome” among the environmental E. coli isolates.

The phylogroups result revealed that the majority of the isolates belonged to the commensal phylogroups A, B1, and C irrespective of sources or places of sampling. Only 2 isolates recovered from HWW sources of Varanasi belonged to the extra-intestinal virulence-associated phylogroup B2. It is documented that, human commensal strains belong mostly to phylogroups A and B1 (Stoppe et al., 2017) and strains isolated from animals fall under group B1 (Higgins et al., 2007). The presence of phylogroup A and B1 may result from anthropogenic activities (Escobar-Páramo et al., 2006; Hanna et al., 2020); this might be the explanation for the result observed in our study. It has been reported that E. coli with extended-spectrum cephalosporin resistance belonged predominantly to non-B2 groups, which was also well observed in our study. Extremely low abundances of phylogroup B2 were reported from Indian aquatic environments (Bajaj et al., 2015) and wastewater contaminated with human and animal waste in Spain (Garcia-Aljaro et al., 2009; Sabaté et al., 2008). Also, it was reported by several researchers that the pandemic ST131 clone of E. coli belonging to phylogroup B2 was often associated with CTX-M mediated ESBL phenotype (Dhanji et al. 2011; Bajaj et al., 2015), but other sequence types belonging to phylogroup B2 have been associated with less CTX-M production (Brisse et al., 2012). In our study, both the phylogroup B2 isolates carried only blaTEM as the major ESBL determinant. Although we could not perform the sequence typing analysis, it may be presumed that a lower prevalence of pathogenic phylogroup B2 in the environmental sources would have a less significant role in the transmission of the pandemic ST131 clones carrying CTX-M genes in those areas. Also, five isolates in this study were found to be grouped under the extra-intestinal pathogenic phylogroup D, with 80% of these harboring the blaCTX-M-gr-1 gene. A similar observation was reported for the E. coli isolates recovered from Yamuna river water in Delhi, India (Bajaj et al., 2015). The presence of the same phylogroups among the E. coli isolates recovered from various environmental sources of a particular area does not necessarily indicate a zoonotic transfer, but rather a common source of acquisition.

5 Conclusion

This study reports the presence of multi-drug resistant ESBL-producing E. coli isolates in wastewater from hospitals, livestock farms, and natural water samples collected from the eastern parts of Uttar Pradesh, India. This is the first research data on the molecular detection of ESBL-producing E. coli isolates involving the three types of environmental sources of the concerned study area. In this study, significant differences in the occurrence of phenotypic ESBL production among the cefotaxime-resistant E. coli isolates recovered from different places (districts) and sources (LFWW, HWS, and NWS) were observed. The blaCTX-M-gr-1 and blaTEM genes were found to be the predominant ESBL genotypes among the recovered isolates. Additionally, pathogenic phylogroups B2 and D were observed to a lesser extent, and a majority of the tested isolates belonged to the commensal phylogroups A, B1, and C. Moreover, the observation of a high degree of resistance to critically important antibiotics among the environmental E. coli isolates of the study area is a matter of serious concern to public health. Further, the use of molecular typing tools will aid in determining the origin and distribution of AMR organisms in a particular geographical area. Optimizing the use of antibiotics in clinical and animal husbandry practices, adopting effective wastewater treatment facilities and routine AMR surveillance among environmental isolates will help to protect public health to a great extent.