The Enterobacterales are a group of ubiquitous Gram-negative bacteria of increasing concern since they are human pathogens and can also transfer AMR genes to other pathogenic bacteria of the same and differing species via MGEs (Meek et al., 2015; Ruh et al., 2019; Arnold et al., 2022). Notably, transmissible resistance to the third generation cephalosporins has become a significant issue due to mounting levels of resistance in hospitals and community settings, which in turn can lead to severe infection and death (Jeanvoine et al., 2020).

WHO recognized and prioritized AMR bacteria into critical, high and medium based on the urgency and need for new antibiotics (WHO, 2019). Ireland, like many other countries has adopted a one health approach to address the challenge of AMR and published the second revision of Ireland’s National Action Plan on AMR in 2021 (Government of Ireland, 2021).

Although limited attention has been given to AMR in isolates from the environment in the past (Singer et al., 2016; Ramovic et al., 2020), EFSA recently produced a scientific opinion on the role played by the environment in the emergence and spread of AMR through the food chain, which identified a large number of data gaps in relation to the sources of contamination, relevance of transmission routes and effectiveness of mitigation measures (EFSA Biological Hazards (BIOHAZ) panel, 2021). It was concluded that there is insufficient data available to support a specific assessment of the quantitative impact that the production environment has on the contamination of foods or on public health due to limited studies on the efficiency of mitigation options for resistant organisms and elimination of resistance genes (EFSA Biological Hazards (BIOHAZ) panel et al., 2021). Antimicrobial resistant organisms (AROs) and MGEs can spread through several different pathways in the environment and have been previously reported in surface waters, soils, animal and human wastewater and foods (Graham et al., 2019). Molecular mechanisms for resistance can be either chromosomal or plasmid mediated, and this impacts on the transmissibility of the genes. The spread of AMR via MGEs such as plasmids, insertion sequences, integrons and transposons plays an important role in antibiotic resistance (Partridge et al., 2018). Among the molecular mechanisms for resistance to extended spectrum cephalosporins (alone or in various combinations) are the presence of genes encoding for TEM, ACT, CMY, SHV, OXA and CTX-M type β-lactamases (De Angelis et al., 2020; Hussain et al., 2021). Molecular mechanisms for chromosomal fluoroquinolone resistance involve mutations in target genes of the quinolone resistance-determining regions (QRDR) such as DNA gyrase (gyrA) or topoisomerase IV (parC or parE) or acquisition of plasmid mediated resistance such as the qnr gene and oqxAB, a quinolone efflux pump (Perez et al., 2013; Azargun et al., 2020; Chen et al., 2020). Characterization of AMR genes by PCR, WGS and a combination of PCR and metagenomics has previously been successfully applied to E. coli isolates from hospital wastewater, grassland soils and fecal samples, manure pits on dairy farms and from the environment and feces on broiler farms (Chandran et al., 2014; Yang et al., 2020; Massé et al., 2021; Byrne et al., 2022). The study of Massé et al. (2021) examined E. coli isolated from fecal samples from calves, cows and manure pits at 101 diary farms in Canada and reported that 85% of farms had at least one AmpC and ESBL producing E. coli and that bla_CMY-2 and bla_CTX-M were the genes, respectively, responsible for these phenotypes. The study of Byrne et al. (2022) investigated the prevalence and frequency of ESBL/AmpC producing E. coli isolated from the environment and fecal samples from broiler farms in Ireland and reported that 13% of E. coli isolates from broiler farms harbored the bla_CMY-2 gene and that hatcheries may be a reservoir and major contributor to the transmission of AMR.

Animal waste can be an important source of AMR bacteria entering the environment, especially if spread on soils where resistance genes can accumulate (Xu et al., 2020; Yang et al., 2020). We recently reported the identification of critically important Enterobacterales in wastewater from pig, cattle and poultry farms, both before and after treatment using ICWs (Prendergast et al., 2022a). Here, we report on the characterization of the isolates collected from farm wastewater before and after ICW treatment, AMR genes, pathotypes, phylogroups, plasmid replicons and virulotypes, and on the relatedness of the E. coli isolates, in order to enhance our understanding of the significance of farm effluents and ICW-treated water as a source of AROs to the environment in Ireland.

The de novo assemblies consisted of an average of 142 contigs (range 56–373 bp) with an average N50 of 162,544 bp (range 66,468–406,961 bp). The average coverage was 88X (range 40–135 X) and core % ranged from 98.8–100 with a mean core percent of 99.8%.

Dissemination of AMR genes in the Irish environment has been largely unknown to date. Our previous study reported the presence of AROs in farm effluents and how ICWs were able to reduce the load of AROs in farm effluent. In this paper WGS results confirmed the mechanisms of resistance to cephalosporins and fluoroquinolones identified phenotypically in most of those isolates. Regarding class A β-lactamase genes, 28 of 30 (93%) presumptive ESBL genomes possessed at least one CTX-M type β-lactamase encoding gene, with bla_CTX-M-14 and bla_CTX-M-15 most prevalent. These results are significant considering the most frequently encountered CTX-M variants in clinical isolates in Europe are CTX-M-14 and CTX-M-15 (Livermore and Hawkey, 2005; Canton et al., 2012; Irrgang et al., 2017; Sarno et al., 2021). The isolates encoding these bla_CTX-M genes were isolated from both farm effluent and ICW treated water, with bla_CTX-M-14 mostly associated with the piggery (15/19) and bla_CTX-M-15 with the dairy farm (5/9).

We identified six presumptive E. coli AmpC producers and all of them carried the bla_CMY-2 gene, five of which were recovered from the pig farm and one from the beef farm. The increase in prevalence of bla_CMY-2 conferring resistance to ceftiofur in pigs receiving a feed medicated with chlortetracycline and penicillin has been shown previously (Jahanbakhsh et al., 2015) and a recent study suggests that plasmid mediated rather than clonal spread likely play an important role for the emergence and transmission of bla_CMY-2 between animals and humans (Ewers et al., 2021). While no E. coli isolates were identified as harboring the bla_CMY-2 gene in the 22 isolates from farm 3, a recent study in Ireland reported that 13% of E. coli isolates from broiler farms harbored this gene (Byrne et al., 2022). While it was reassuring that no carbapenem resistance genes were found in any isolate in this study, it has been reported that AmpC enzymes promote carbapenem resistance in isolates with ESBL or defects in permeability (Mirsalehian et al., 2014; van Boxtel et al., 2016).

In the livestock and animal sectors, CTX-M-1, CTX-M-14 and CTX-M-15 ESBL types have frequently been detected in a number of different countries in Europe (Day et al., 2016; Ewers et al., 2021), as well as the United States (Afema et al., 2018), Canada (Cormier et al., 2019), and New Zealand (Collis et al., 2022). Agersø et al. (2011) also reported bla_CTX-M-14 and bla_CTX-M-15 in Danish pigs and pork with the highest prevalence was for bla_CTX-M-1, in contrast with our results. The isolates from the study of Agersø et al. were further characterized (Hammerum et al., 2012) and they reported that pigs and pork can be a reservoir of ExPEC CTX-M-14-producing E. coli. We also identified ExPEC carrying bla_CTX-M-14 in this study. Ewers et al. (2021) conducted a large-scale surveillance which included nine European countries, but not Ireland, where 2,993 commensal Escherichia spp. isolates were recovered from randomly collected fecal samples of healthy cattle, pigs and chickens in various abattoirs. By analyzing 99 isolates using WGS they identified bla_SHV–12 (32.3%), bla_CTX–M–1 (24.2%), and bla_CMY–2 (22.2%) as the predominant ESBL/pAmpC types (Ewers et al., 2021). It is generally accepted that it is very difficult to compare countries and farms unless all studies are conducted the same, i.e., variations such as sample size, animal age, health status, variation in farming systems between countries, sample matrices and culture selection methods are all important variations to consider when comparing studies (Collis et al., 2022). This is why EFSA has recommended a joint European approach in order to address this knowledge gap (EFSA Biological Hazards (BIOHAZ) panel et al., 2021).

Although only encountered in one sample, the bla_CTX-M-65 carrying E. coli isolate is of interest since it was from an ICW-treated water sample, and therefore it had the potential to enter waterways and the environment. CTX-M-65 has been previously referred to as a hybrid enzyme of both CTX-M-14 and CTX-M-15 with enhanced ESBL activity (He et al., 2015), and our isolate, in addition, also harbored qnrS1. A study in Bolivia showed that in the absence of selective pressure from antimicrobials, plasmid transfer of bla_CTX-M-65 to other pathogens occurs frequently and is stable (Riccobono et al., 2015). In Europe, bla_CTX-M-65 has also been reported in S. Infantis isolated from broilers and humans in Italy (Franco et al., 2015), in E. coli from cattle in the Netherlands (Dantas Palmeira and Ferreira, 2020) and more recently from beef and pork samples collected at retail in Portugal (Leão et al., 2021) and in S. Infantis from human cases in Spain (Vázquez et al., 2022). To the best of the authors’ knowledge, this is the first reported case of a bla_CTX-M-65 isolated from the environment in Ireland.

While no carbapenemase or colistin resistant isolates were identified phenotypically, all isolates were screened for carbapenemase and colistin resistance genes and none were identified with the exception of the multidrug resistant (MDR) presumptive AmpC producing, and fluoroquinolone resistant E. cloacae isolated from the piggery that harbored an mcr-9 gene (data not shown) in addition to bla_ACT-14, bla_SHV-12, bla_TEM-1B and qnrB2. The dissemination of MDR E. cloacae throughout the UK and Ireland has been previously reported and is now recognized as the third major nosocomial infection after E. coli and K. pneumoniae (Moradigaravand et al., 2016). The bla_ACT-14 gene has been previously observed in E. cloacae isolates from dogs and wild birds (Literak et al., 2014; Boehmer et al., 2018). Our isolate also harbored the IncHI2-ST1 plasmid replicon; the family of IncHI2 plasmids often carry MDR genes that can be transferred horizontally along the food chain (Diaconu et al., 2021).

Plasmids encoding ESBLs and pAmpCs frequently harbor additional resistance genes and so can present a significant therapeutic challenge (Gupta and Bhadelia, 2014) and since some groups of antibiotics may persist in the environment for a very long time, i.e., fluoroquinolones have the lowest rate of degradation and highest resistance forming potential within the environment (Department of Health (DOH), 2017), it would be of interest to determine if the use of fluoroquinolones at farm level could increase the risk or persistence of resistant organisms and/or their resistance genes. Unfortunately, we did not have access to information on antimicrobial usage on any of the farms for this study. Lack of knowledge of antimicrobial usage at farm levels in Ireland has been previously highlighted (Martin et al., 2020) but will hopefully be the focus of future studies in our laboratory where we aim to examine the relationship between the use of on farm antibiotics and AMR. However, studies in the Netherlands and New Zealand (Hordijk et al., 2019; Collis et al., 2022) reported that antimicrobial usage only partially explained ESBL and AmpC positive samples and highlighted that factors other than antimicrobial usage may contribute to the transmission of AMR in the farm environment.

WGS also confirmed the phenotypic fluoroquinolone resistance observed and whether it was chromosomal, or plasmid mediated. While the presence of plasmid mediated genes has an additive effect and pose a higher risk of AMR dissemination, the low numbers of qnr genes in comparison to chromosomal resistance observed in this study is in agreement with other studies (Jacoby et al., 2003, 2006; Strahilevitz et al., 2009). Among isolates phenotypically resistant to CIP only and considered therefore presumptive plasmid mediated quinolone resistant isolates (Prendergast et al., 2022a), the most common gene found was qnrS1, present in seven isolates compared to qnrB1, qnrB2 and oqxB, which were each encountered in one instance only. Similar results were also observed in a study that reported qnrS followed by qnrB to be the most prevalent genes in E. coli from livestock and food in Germany (Juraschek et al., 2021). The oqxB gene was only observed in our study in the K. pneumoniae isolate phenotypically described as an ESBL producer susceptible to quinolones. While both qnr and oqx genes are plasmid mediated, oqx genes are members of the resistance-nodulation-division (RDN) efflux pump that have been shown to confer resistance to numerous antibiotics in addition to quinolones, including nitrofurantoin, quinoxalines, tigecycline, chloramphenicol, detergents, and disinfectants (Bharatham et al., 2021) and their presence has been shown to accelerate the development of fluoroquinolone resistance in S. Typhimurium (Wong et al., 2014).

All of the 47 presumptive chromosomal fluoroquinolone resistant isolates (resistant to both CIP and NAL) had multiple mutations in gyrA and/or parC, which have been associated with the expression of high-level fluoroquinolone resistance in clinical isolates (Ruiz et al., 1995; Minarini and Darini, 2012). In our previous study the MIC values were high, i.e., all 47 isolates displayed MIC values >128 μg/mL for NAL and ≥ 8 μg/mL for CIP. About one quarter of all isolates showed co-resistance to cephalosporins plus fluoroquinolones, which can be attributable to chromosomal mutation or acquisition of genes by horizontal transfer as explained by Tacão et al. (2014). Other researchers have shown that qnr genes can co-exist with the bla_TEM and bla_SHV alleles in clinical isolates and is most likely plasmid mediated (Jiang et al., 2008; Mood et al., 2015). Plasmid mediated resistance was observed in two isolates in our study with E. cloacae isolated from farm 4 effluent harboring bla_SHV-12, bla_TEM-1B and qnrB2 (data not shown for TEM) and an E. coli isolate from farm 1 effluent that harbored bla_SHV-12, bla_TEM-1B and qnrS1.

Amongst the 10 E. coli isolated from ICW-treated water, we identified bla_CTX-M-14 in one isolate, qnrS and bla_CTX-M-65 in another isolate and three isolates carried mutations in the AmpC promoter region in addition to mutations in gyrA and parC. In addition, double mutations in gyrA were identified in five of the isolates as well as single or double mutations in parC. The E. coli harboring bla_CTX-M-14 was of predicted pathotype UPEC and one E. coli that carried a mutation in the AmpC promotor in addition to double mutations in gyrA and a single mutation in parC was of predicted pathotype ExPEC. The finding of AMR Enterobacterales along with MGEs from the ICW-treated water is concerning considering wildlife, arthropods, air and dust can act as vectors for short and long-distance environmental transmission of AMR contamination (EFSA Biological Hazards (BIOHAZ) panel et al., 2021). Resistant organisms and ARGs can enter groundwater through infiltration or bank filtration from surface water or leaching from the soil. This is of additional concern considering ground water is frequently used as a source of drinking water (Felis et al., 2020) and could also be absorbed by food crops, further contributing to exposure to and persistence of these contaminants (Department of Health (DOH), 2017).

The majority (54%) of our E. coli isolates grouped into phylogroup A, which was previously reported as the group where most human commensal strains worldwide belong to (Stoppe et al., 2017; Newman et al., 2021). The second most common phylogroup in our study was B1 (28%). No clear association between farms and phylogroups could be concluded, except for the isolates represented by phylogroup E which were confined to poultry, and this agrees with the study of Logue et al. (2017). Differences in phylogroup distribution have been observed previously and may be attributed to different circumstances such as geographical and climate location. A higher frequency of groups A and D from raw wastewater samples in Brazil was previously reported (Stoppe et al., 2017) whereas other studies of waste, ground and recreational waters reported higher frequencies of groups B2 followed by D in Australia (Anastasi et al., 2010), D followed by A in Poland (Mokracka et al., 2011) and A followed by B1 in Portugal (Figueira et al., 2011), the latter much the same as our results.

All five isolates identified as UPEC were grouped into phylogroup D and isolates within this group are considered of risk to public health (Orsi et al., 2007; Vázquez-Villanueva et al., 2023). While other studies reported that the predominant phylogroup in UPEC isolated from Europe and elsewhere is B2, followed by D (Dubois et al., 2010; Ejrnæs et al., 2011; Halaji et al., 2022), no isolate of phylogroup B2 was identified in this study. Since E. coli ST131 is considered an important emerging pathogen among phylogroup B2 strains harboring multiple AMR and virulence genes (Halaji et al., 2022), it was reassuring to note that there was no isolate of phylogroup B2 or ST131 identified in this study.

Isolates identified as ExPEC were not confined to one phylogroup, i.e., groups A (7 isolates), B1 (7 isolates) and C (four isolates). These findings therefore do not fully agree with Dale and Woodford (2015) and Sarowska et al. (2019), that indicated that the E. coli responsible for intestinal infections represent phylogenetic groups A, B1 or D and extraintestinal infections represent groups B2 and D.

Analysis of the MLST data revealed an association of certain STs with different farms, i.e., ST44 and ST48 were commonly associated with the piggery, ST58, ST156 and ST4385 with the beef farm, ST744 with the dairy and poultry farm and the two isolates recovered from the dairy farm were ST720. E. coli ST10 is recognized as an important human pathogen (Manges and Johnson, 2012) and have been previously isolated from poultry and pig sources (Bergeron et al., 2012). Among the 20 ST10 isolates in this study, 7 were of predicted ExPEC pathotype and this pathotype was also identified among E. coli isolates of ST88, ST162, ST58 and ST117. When the distribution of the major STs (ST10 and ST744) were compared between poultry and non-poultry farms, neither STs appeared to be overrepresented in a particular type of farms, although there was a larger proportion of ST744 among poultry vs. non-poultry farms than ST10. However, due to small numbers of isolates, i.e., 20 isolates of ST10 (3 poultry; 17 non-poultry) and nine isolates of ST744 (3 poultry and 6 non-poultry) it is difficult to make definitive conclusions, and this should be the focus of future studies.

ExPEC pathotypes were identified in all four farm types with the highest number identified in the beef farm (7 isolates) and the lowest in the dairy farm (3 isolates). This is concerning since ExPEC is a leading cause of urinary tract infections and responsible for the death of thousands of people every year especially young children (Munhoz et al., 2021). UPEC is the most common pathotype causing urinary tract infections (Guglietta, 2017; Rezatofighi et al., 2021; Hasan et al., 2022) and this pathotype was identified in six isolates and in three of the four farms with the highest number in the beef farm (3 isolates). ExPEC and UPEC can easily acquire MGEs and virulence factors from related bacteria and usually contain multiple pathogenicity islands (Johnson and Russo, 2005). Of interest to note was a ST117 isolate that was identified as a hybrid (ExPEC/UPEC) E. coli strain from the dairy and poultry farm which warrants further investigation regarding pathogenicity as they are considered to be more virulent (Santos et al., 2020; Munhoz et al., 2021). Of additional concern is that this isolate was grouped into phylogroup G. Isolates of this phylogroup have been identified as a poultry lineage with high virulence and antibiotic resistance potential and previous epidemiological data on 4,524 French and Australian strains suggested that phylogroup G of ST117 can also establish in humans and cause extra-intestinal diseases (Clermont et al., 2019). The APEC pathotype (presence of three or more of the genes iss, iroN, hlyF, ompT, and uitA) was identified in 17 isolates with the majority identified at farm 4, piggery, (five isolates) with four isolates identified as APEC at each of the other three farms.

Plasmids play a critical role in the mobilization of AMR genes between bacteria and are one of the main reasons for resistance among Gram-negative bacteria (Li et al., 2019). In our study differences in plasmid replicon types in E. coli isolates from the same cluster that originated from different farms or from different sampling dates suggests local/regional transmission of plasmid replicons. We were however unable to confirm with the short-read sequencing that was used in this study if resistance genes were transferred by particular plasmids. More research is needed to decipher the role of particular plasmids in the transmission of AMR in the environment by undertaking full length sequencing of plasmids, and to better understand the ExPEC population and its role in AMR in the Irish farm environment.

More than 50 genes encoding virulence determinants, were identified, at different frequencies; terC encodes a protein that functions in tellurium ion resistance and was identified in all 79 E. coli isolates, whereas other virulence genes were present in only one isolate. These virulence genes were associated with functions including stress, survival, regulation, iron uptake secretion systems, invasion, adherence and toxin production. The abundance of the tellurium ion resistance gene was surprising, as other studies have not reported such a high prevalence. A recent study of 167 E. coli genomes from human and animal environments circulating in 40 Brazilian cities identified all isolates as negative for terC (Fuga et al., 2022). Other authors found that tellurium ion genes are mostly associated with STEC belonging to serotypes such as O26, O103, O111 O121, O145, and O45 (Lewis et al., 2018; Nguyen et al., 2021) none of which were isolated in our study (serotype data not shown). The gene traT whose protein functions in outer membrane protein complement resistance, was the next most common virulence gene and was identified in 71% of isolates. It has been suggested that this gene may have a role in the pathogenesis of mastitis (Agüero et al., 1984), cystitis (Hasan et al., 2022) and pyelonephritis (Firoozeh et al., 2014). The serum survival gene iss and the sitA gene responsible for iron uptake, were identified in 46.7 and 43% of E. coli isolates, respectively, and they have been previously reported to be highly prevalent among avian pathogenic E. coli (Johnson et al., 2008; De Carli et al., 2015; Lima et al., 2019). In this study both of these genes were identified in isolates from all four farms without significant differences among them. In terms of pathogenicity, isolates identified as ExPEC, UPEC or hybrid (ExPEC and UPEC) had a larger prevalence of sitA and iss genes (alone or in combination) only in the beef farm (identified in 78.6% of isolates from the beef farm and 44.4, 40 and 41.6% at the dairy, dairy and poultry and pig farms respectively). No other association was observed in this study between STs, pathotypes, AMR types or species pathotypes and virulence genes.

In conclusion, to the best of the authors’ knowledge, this is the first study to provide insights into the prevalence of ESBL, AmpC and qnr genes in Enterobacterales collected from untreated and treated farm wastewater in Ireland. Our study shows that low levels of antimicrobial resistant Enterobacterales, persist even following wastewater treatment and that they are equipped with plasmids to transmit resistance genes of clinical relevance to the environment. This could ultimately lead to the contamination of drinking water or irrigation water used in crops. Further investigation on the role of plasmids in multiple antibiotic transfer mechanisms will be the focus of future studies in our laboratory.

Sequencing data generated for this study were deposited to the NCBI Sequence Read Archive (SRA) repository and can be found through the BioProject accession number PRJNA862318 https://www.ncbi.nlm.nih.gov/bioproject/PRJNA862318.

DP, CB, DMo, and MG conceived the study. DMu and DJ performed WGS and uploaded sequences to NCBI. ÁO’D and JM contributed to provision of materials. DP analyzed data and wrote the manuscript with support from MG and RS. All authors contributed to the article and approved the submitted version.

This work was carried out as part of the AREST project which is jointly funded by the Environmental Protection Agency, under the EPA Research Programme 2014–2020, and the Health Service Executive (2017-HW-LS-1). The EPA Research Programme is a Government of Ireland initiative funded by the Department of Communications, Climate Action and Environment. It is administered by the Environmental Protection Agency, which has the statutory function of co-ordinating and promoting environmental research.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.