Antimicrobial resistance and phylogenetic relatedness of extended-spectrum ß-lactamase (ESBL)-producing Escherichia coli in peridomestic rats (Rattus norvegicus and Rattus tanezumi) linked to city areas and animal farms in Hong Kong

. Antimicrobial resistance and phylogenetic relatedness of extended-spectrum ß-lactamase (ESBL)-producing Escherichia coli in peridomestic rats (Rattus norvegicus and Rattus tanezumi) linked to city areas and animal farms in Hong Kong


Introduction
Escherichia (E.) coli is a typical commensal resident of the mucus layer of the vertebrate colon and as such rarely causes disease (Kaper et al., 2004).E. coli is also highly adaptable to other ecological niches and is a versatile pathogen associated with a range of intestinal and extra-intestinal diseases and with antimicrobial resistance (AMR) (Foster-Nyarko and Pallen, 2022).AMR is a significant threat to human and animal health worldwide, since resistant pathogens are more difficult or even impossible to treat (WHO, 2019).One of the most important mechanisms of AMR in E. coli and other Gram-negative bacteria is the enzymatic inactivation of penicillins and cephalosporins by extended-spectrum β-lactamases (ESBLs), such as the TEM-, SHV-, or CTX-M enzymes (Bush, 2013).Extended-spectrum β-lactamases-producing E. coli (ESBL-EC) are of concern, since they are not only resistant to 3rd generation cephalosporins which are ranked by the World Health Organization (WHO) as critically important antimicrobials for human medicine (Aidara-Kane et al., 2018), but often express a multidrug-resistant (MDR) phenotype, leaving only limited treatment options (Pitout and Laupland, 2008).
In the last decades, ESBL-EC have become prevalent within and beyond healthcare settings, in livestock and in companion animals (Peirano and Pitout, 2019).Furthermore, several studies have shown that multidrug resistant, highly virulent human-related clonal lineages of extra-intestinal pathogenic E. coli (ExPEC) belonging to sequence type (ST) 38, ST58, ST95, and ST131 may be isolated from urban wildlife, including birds and rodents (Guenther et al., 2012;Schaufler et al., 2018;Desvars-Larrive et al., 2019).
Rattus spp.Are widespread pest rodents in urban areas and on livestock farms and may therefore be considered very good sentinels for the occurrence of AMR and ESBL-EC in their surroundings (Dominguez et al., 2023;Strand and Lundkvist, 2019).Antimicrobial-resistant Enterobacterales were isolated previously from urban rats in Berlin, Germany (Guenther et al., 2012), Conakry, Guinea, West Africa (Schaufler et al., 2018).Guadeloupe, French West Indies (Guyomard-Rabenirina et al., 2020), Hanoi, Vitenam (Le Huy et al., 2020), Hong Kong (Ho et al., 2015), Makoku, Gabon, Central Africa (Onanga et al., 2020), Piraeus, Greece (Burriel et al., 2008), Vancouver, Canada (Himsworth et al., 2015) and Vienna, Austria, (Desvars-Larrive et al., 2019), while Nhung et al. (2015) and Dominguez et al. (2023) described AMR Enterobacterales in farm rats in Vietnam and Buenos Aires province, Argentina, respectively.However, recent data on the prevalence of ESBL-EC in peridomestic Rattus spp.are still scarce, and there is currently a lack of information on the spatial distribution and variability of commensal or clinically relevant rat-borne ESBL-EC in different environments (Uea-Anuwong et al., 2023).Therefore, this study was designed to evaluate the occurrence of ESBL-EC in peridomestic rats captured in different city areas and livestock farms in Hong Kong during 2020-2021 and to characterize the isolates using phenotypic and genotypic methods including whole genome sequencing analyses.
Emphasis was placed on identifying bla ESBL and other AMR genes, and on assessing the phylogenetic relatedness of ESBL-EC obtained from rats trapped in different locations.By using this approach, the study aimed to provide insights into the potential role of rats in the presence and spread of AMR in urbanised and livestock-related environments.

Sampling
Rats were trapped during October 2020-August 2021 using live trap cages containing baits (bacon, peanut butter and dried fruit).Each cage was numbered, and the locations were recorded (Table S1).Geographical mapping was performed using R package version 2.11.2 (https:// github.com/r-spatial/mapview)(Appelhans et al., 2023).A total of 16 locations (designated A-P) in nine districts of Hong Kong were selected which included 10 locations within residential areas (city areas), two chicken farms (chicken farm I and chicken farm II, respectively), two pig farms (pig farm I and pig farm II, respectively), and two stables belonging to horse-riding schools (horse-riding school I and horse-riding school II, respectively; Fig. 1).Baited cages were checked every morning.Each trap containing a rat was placed into a hermetic container.Euthanasia was induced by placing a conical tube containing cotton pads soaked in 5% isoflurane into the container with the animal, and sealing the lid, following the guidelines of the American Veterinary Medical Association (Underwood and Raymond, 2020).The collected rats were taken to the Veterinary Diagnostic Laboratory, City University of Hong Kong for further investigation.After removal of the rats, traps were cleaned out and submerged in broad spectrum surface disinfectant (Rely + On Virkon™, Sudbury, UK) for 10-15 min, brushed, and rinsed in cold water prior to reuse.
Post-mortem examinations of the rats were performed by a certified  veterinarian.In the case of a rat responding to stimulus (toe pinch), cervical dislocation was performed as an additional method of euthanasia, according to the guidelines of the American Veterinary Medical Association (Underwood and Raymond, 2020).During post-mortem examinations, caeca were aseptically collected, placed into sterile 2 mL Eppendorf tubes and processed on the same day.

Bacterial isolation and identification of ESBL producers
Caecal swabs were enriched in 2.7 mL brain heart infusion (BHI) broth (Becton Dickinson, Shanghai, China) with cefotaxime (3.5 mg/ mL) and vancomycin (32 mg/mL) at 37 • C for 24 h.One loopful of the broth was streaked onto Brilliance extended-spectrum ß-lactamase (ESBL) agar (Thermo Fisher Scientific, Melbourne, Australia) and incubated at 37 • C for 24 h.Putative ESBL producing E. coli were selected based on their colours and morphologies and subjected to species identification by Matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDITOF; Bruker, MA, US), and the MALDI Biotyper (Bruker, MA, US).For calibration and internal quality control of the MALDI Biotyper System, the Bruker Bacterial Test Standard (BTS) containing extract of E. coli DH5α was used, according to the manufacturer's instructions (Bruker).Extended-spectrum ß-lactamase production was confirmed using the Combination Disk Test and cefotaxime and ceftazidime with or without clavulanic acid, following the Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI, 2023).

Rat species identification
The genomic DNA of rats was extracted from liver tissue and the species of rat were identified based on the method described by Huang et al. (2022).

Ethics statement
This study was approved of by the Animal Research Ethics Sub-Committee of the City University of Hong Kong (Internal Ref: A-0380).

Rodent capture and location characteristics
A total of 229 rats were captured, whereof 221 were available for the study.Four rat species were identified including R. norvegicus (n = 144), R. tanezumi (n = 67), R. andamanensis (n = 8), and N. huang (n = 2).Thereof, 104 were from residential areas, 44 originated from chicken farms, 52 from pig farms, and 21 were captured in the stables of horseriding schools (Table 1).

Occurrence of ESBL-producing E. coli among Rattus spp.
E. coli displaying the ESBL phenotype were isolated from the caecal samples of 59% (130/221) of the rats.The rat species included R. norvegicus (n = 110), and R. tanezumi (n = 20).No ESBL-producing E. coli were isolated from R. andamanensis and N. huang.ESBLproducing E. coli were isolated from 56% (58/104) of the rats trapped in residential areas, from 80% (35/44) of the rats from chicken farms, from 69% (36/52) of the rats caught on pig farms, and from 5% (1/21) of the rats trapped on the premises of horse-riding schools.One E. coli isolate per rat sample was retrieved, except for four rats (R097, R107, and R117, from city areas, and R065 from chicken farm II, respectively) which yielded two distinct isolates each (Table 1).In total, 134 E. coli were available for further characterisation.

Antimicrobial resistance phenotypes and genotypes of rodent ESBL-EC
Antimicrobial susceptibility test revealed resistance to the 3rd generation cephalosporin cefotaxime in all but one E. coli (isolate HK127 originating from a city area) which was classified as intermediately

Location of bla ESBL genes and similarity of bla ESBL -containing plasmidal contigs
In our assemblies obtained from short-read sequencing data, most bla CTX-M genes were situated on short contigs (median length 4.9 kb), precluding further investigations regarding their genetic context or potential horizontal transmissions among the rat isolates.The bla CTX-Mcontaining contigs from 42 isolates had a length of >20 kb and were further examined.Analyses with Platon (Schwengers et al., 2020) predicted a plasmidal location of bla CTX-M in 14 of these isolates (Supplementary Table S2).High pairwise similarities (mash distances <0.01) were found for seven of the 14 plasmidal contigs.These comprised contigs (67 and 72 kb; bla CTX-M-14 , IncB/O/K/Z) from two ST58 isolates (HK6 and HK8 from rats trapped in the city), contigs (24 and 26 kb; bla CTX-M-15 , no replicon identified) from two ST10 isolates (HK17 and HK26 from chicken farms), and contigs (35-41 kb; bla CTX-M-55 , IncX1) with from three unrelated isolates HK74 HK101, HK121, (ST1112, ST1286, and ST1011, respectively), all linked to city areas.The ST58 and ST10 isolates each belonged to SNP clusters, indicating vertically inherited plasmids.The similar contigs of the three unrelated isolates may represent a segment of a circulating bla CTX-M-55 IncX1 plasmid.

Phylogenetic relatedness of rodent ESBL-EC from different locations
To determine potential transmission clusters, isolates sharing a sequence type were further compared using cgSNP.A total of 17 distinct clusters of two or more identical (<10 cgSNP distance) isolates was identified.ESBL-EC from the two largest clusters (nine and seven isolates, respectively) were obtained from rats trapped on pig farm II and included all nine E. coli ST1081, and seven E. coli ST44, respectively (cluster 8 and cluster 17, respectively in Fig. 5).A cluster of five E. coli ST155 (cluster 3 in Fig. 5) originated from chicken farm II and were unrelated by cgSNP to the remaining E. coli ST155 from the same farm (n = 4) and from a residential area (n = 1).Further, a total of five E. coli ST5853, all collected from rats originating from chicken farm II belonged to two separate clusters consisting of three and two isolates, respectively (cluster 2 and cluster 1, respectively in Fig. 5).Among the 10 E. coli ST14801 isolated from rats from pig farm II, two clusters (cluster 12 and cluster 13, respectively) containing three isolates each were determined, and the remaining four E. coli ST14801 differed by 10-26 pairwise cgSNPs from the clustered isolates.Similarly, a subset of the seven E. coli ST58 retrieved from rats trapped within city areas belonged to two different clusters consisting of two isolates each (cluster 4 and cluster 11 in Fig. 5), and the remaining isolates (n = 3) were unrelated by cgSNP.Just two E. coli belonging to the ST10 collection (n = 9) were related by cgSNP and originated from chicken farms I and II (cluster 6 in Fig. 5).
The remaining seven clusters (5, 7, 9,10, and 14-16) comprised two to three isolates each and spanned collection points from residential areas, chicken farm II, and pig farm II (Fig. 5).Except for cluster 6, all clusters observed in this study were linked to the trapping locations of the rats.Identical bla ESBL genes and similar genotypic and phenotypic AMR patterns were found for most of the clustered isolates, supporting their close genetic relatedness (Fig. 5).

Comparison of rodent ESBL-EC with globally occurring E. coli isolates
To investigate our isolates in the context of global E. coli, read data of the 134 rat isolates were submitted to Enterobase (Achtman et al., 2022).Putative transmission clusters (<10 SNPs) determined for our Hong Kong rat isolates using the CFSAN pipeline corresponded to an Enterobase hierarchical clustering (HierCC cgMLST) at the HC2 or HC5 levels (pairwise distances ≤2 or ≤5 alleles).None of our rat isolates clustered with any of the 277,500 publicly available E. coli genomes at HC5 level.At HC10 level (pairwise distances ≤10 alleles), two clusters involving the Hong Kong rat and global isolates were observed: rat isolate HK115 (ST744; trapped in the city area) clustered with a pig  manure isolate obtained in 2015 in Vietnam and a second isolate with missing metadata.Further, rat isolate HK71 (ST648; trapped in the city area) clustered at HC10 level with 10 global isolates from human (n = 6), water (n = 3), or unknown (n = 1) samples collected between 2017 and 2021 predominantly in Asia including China, Japan, and Thailand (n = 7), but including isolates from Europe and Australia.The genomes of HK115 and HK71 belonged to HC10 cluster number 126191 and HC10 cluster number 102528, respectively (http://enterobase.warwick.ac.uk).These metadata suggest that this HC10 cluster reflects a long-term and globally circulating strain rather than direct transmission events.

Discussion
The present study includes an assessment of the occurrence of rat-borne ESBL-EC in Hong Kong during 2020-2021.Notably, ESBL-EC were isolated only from peridomestic rat species (R. norvegicus and R. tanezumi) mostly collected from city areas and animal farms.Wild or non-urban rat species namely R. andamanensis and N. huang (Chung and Corlett, 2006) were only trapped on the premises of the two horse-riding schools and were absent of ESBL-EC.
The overall prevalence of ESBL-EC among the rats investigated in this study was 59% which is considerably higher than the prevalence of 13% reported in an earlier study investigating a collection of 491 brown and black rats captured in Hong Kong during 2008-2013 (Ho et al., 2015).It is also noticeably higher than the ESBL-EC prevalence reported in rats in city areas of Berlin, Germany in 2010 (16% of 56 rats) (Guenther et al., 2013), in Conakry, Guinea, West Africa, in 2015 (10% of 29 rats) (Schaufler et al., 2018)   (6.8% of 144 rats) (Le Huy et al., 2020), and in Guadeloupe, 2013-2014(3.1% of 162 R. norvegicus) (Guyomard-Rabenirina et al., 2020).Notably, variations between countries for which comparative data are available should be interpreted with caution due to the differences in study designs, sample types, and testing methodologies.Further, many studies on AMR in Rattus spp.do not determine the specific genotype of underlying AMR resistance such as the occurrence of bla ESBLs , with many investigations characterizing the antimicrobial resistant bacteria only phenotypically (Uea-Anuwong et al., 2023).Therefore, the data from the present study contribute to current understanding of spatio-temporal variations of rat-borne ESBL-EC and highlight the role of peridomestic rats as reservoirs and spreaders of ESBL-EC (Uea-Anuwong et al., 2023).Further, E. coli in rats are thought to serve as indicators of clinically relevant drug-resistant Enterobacterales, including ESBL-EC that may be circulating in the human population or within livestock herds (de Cock et al., 2023;Strand and Lundkvist, 2019).In this study, the prevalence of ESBL-EC among rats captured in city areas of Hong Kong was 56%, which correlates well with the observed prevalence of asymptomatic faecal carriage of ESBL-Enterobacterales in the Hong Kong human community in 2017 (52.8%) (Kwok et al., 2020).Likewise, the detection of ESBL-EC in rats from chicken farms (80%) and from pig farms (69%) is reflective of the observed prevalence in chicken (78.6%) and pigs (75.4%) in livestock farms in China (Liu et al., 2021).By contrast, the frequency of ESBL-EC among rats collected from horse stables was very low which might mirror low levels of antimicrobial use in horses.Furthermore, the two horse-riding schools were located far away from the city and livestock farms, with one location (location P) being adjacent to a country park.This may explain why all (with one exception) of the two main non-urban species (R. andamanensis and N. huang) were solely caught from these locations.These rodent species thrive in forests, shrubland, and grassland characterised by little anthropogenic activity (Chung and Corlett, 2006).
Although diverse, the predominant bla CTX-M genes were bla CTX-M-14 and bla CTX-M-55, both of which are highly prevalent among Enterobacterales from human and animal populations in China (Bevan et al., 2017).Other prevalent variants included bla CTX-M-65 which is frequently reported among E. coli and Salmonella in Asia (Furlan et al., 2020), and bla CTX-M-15 .The high prevalence of bla CTX-M-15 among rats in this study is of public health concern since it is widespread among ESBL-EC causing human infections (Bevan et al., 2017;Cantón and Coque, 2006).Rats from city areas and chicken farms were shown to carry an abundance of different bla CTX-M variants, compared to rats from other locations.Three of the most prevalent bla CTX-M variants (bla CTX-M-14, bla CTX-M-15 and bla CTX-M-65 ) were shared among isolates from city areas, chicken farms and pig farms, which is in agreement with their wide dissemination in human and animal settings (Palmeira et al., 2021;Bevan et al., 2017).Considering the limitations of short-read sequencing, complete plasmid sequences were not reconstructed, and it was only possible to identify plasmid replicons IncB/O/K/Z and IncX1 which were identified in E. coli linked to city locations.These plasmid replicons have been frequently reported harbouring clinically important resistance genes in food and clinical E. coli genomes deposited in the GenBank database until 2020 (Balbuena-Alonso et al., 2022).
In addition to bla ESBL , a plethora of additional AMR genes were identified among the isolates, irrespective of the origins of the rats.Most of these AMR genes confer resistance to antimicrobials that, besides 3rd and 4th generation cephalosporins, are listed as critically important, or highly important for use in human medicine, such as aminoglycosides, fluoroquinolones, macrolides, sulfonamides and tetracyclines (WHO, 2019).Accordingly, phenotypic resistance to these classes of antimicrobials and MDR was observed frequently.There were some differences in phenotypic resistance traits in that ESBL-EC linked to chicken farms showed a higher rate of resistance to ciprofloxacin than ESBL-EC from pig farms, whereas ESBL-EC from pig farms were more frequently resistant to chloramphenicol and tetracycline than isolates associated with other locations.These observed variations could reflect the applications of different antimicrobials used for treatment, prophylaxis, or growth promotion in chicken and pig farming, respectively, in the study area.Data on antimicrobial usage in livestock in Hong Kong are currently lacking.However, in China, fluoroquinolones including ciprofloxacin are permitted for use in poultry, and amphenicols and tetracyclines are widely applied in swine in China, thus, a similar trend may be expected for Hong Kong (Du et al., 2020;Roth et al., 2019;Yang et al., 2019).
Notably, two ESBL-EC in this study harboured the plasmid-mediated polymyxin E (colistin) resistance genes mcr1.1 (isolate HK 104) and mcr-1.26(isolate HK38), respectively.Since mcr-1 was first reported in China in 2015, mcr variants have been identified worldwide in a wide range of Gram-negative bacteria from human, animal, food, and environmental sources (Nang et al., 2019).Although mcr-1 has been detected in E. coli from urban rats in Hanoi (Le Huy et al., 2020), and rats from a pig farm in South America (Dominguez et al., 2023), to our knowledge, this is the first description of mcr-harbouring ESBL-EC isolated in peridomestic rats captured in city areas and on chicken farms.The presence of mcr genes in Enterobacterales is highly concerning, since colistin has become a last-resort antimicrobial to treat life-threatening infections due to MDR Gram-negative bacteria.(Paterson and Harris, 2016).Likewise, the 16s RNA methylase gene armA, identified in isolate HK68 in this study, represents an emerging and clinically most significant mechanism of high-level pan-aminoglycoside resistance (Yang and Hu, 2022).
The evaluation of the E. coli phylogroups showed that the majority (96/134) belonged to group A and B1.Markedly, several investigations into the worldwide distribution of commensal E. coli in healthy humans showed that while phylogroup A is most common globally, in some countries, E. coli B2 (China, France, and Japan) and E. coli D (Korea) increasingly dominate the commensal E. coli of healthy humans (Stoppe et al., 2017;Massot et al., 2016).By contrast, only a minority (5/134) of the rodent E. coli from this study belonged to phylogroup B2.This suggests that E. coli B2, despite their frequency among humans, may poorly colonise the gut of the urban rat species analysed in this study.This observation should be taken into consideration when assessing rats as sentinels for AMR E. coli in their environments.
Phylogenomic analysis of the ESBL-EC showed that the diversity of STs was highest among isolates associated with rats from city locations, and that few STs were shared between the city locations and farm locations.Several STs associated with ExPEC were detected among isolates linked to city areas, including the pandemic MDR clonal lineage E. coli ST131 and the recently emerged MDR high-risk clone ST1193 (Pitout et al., 2022;Tchesnokova et al., 2019).Both clones are an important cause of bacteraemia and urinary tract infections and play a pivotal role in the global dissemination MDR E. coli (Pitout et al., 2022;Nicolas-Chanoine et al., 2017).Further high-risk international E. coli clones identified in the present study included pandemic uropathogenic E. coli (UPEC) ST69 and E. coli ST73 (Kocsis et al., 2022), as well as E. coli ST38 and E. coli ST648 which are reported from diverse origins, including humans, companion animals, livestock and wildlife (Schaufler et al., 2019;Ewers et al., 2014).E. coli ST58, identified in this study in seven rats, is an emerging human pathogen that is unusual among ExPEC in that it belongs to phylogenetic group B1 (Reid et al., 2022).While the factors that contributed to the emergence and pathogenicity of this lineage remain unclear, our data suggest that peridomestic rats may serve as a non-human source (Reid et al., 2022).
E. coli ST10, and ST155 were the only STs common to ESBL-EC from city areas and chicken farms, and E. coli ST44 the only ST shared between isolates from city areas and pig farms.SNP analysis however revealed that the isolates belonging to common STs fell into distinct SNP clusters, indicating a limited dispersion of strains between the city locations and the animal farms.Interestingly, our observations correlate with recent One Health based investigations of ESBL-EC at the humanlivestock interface which found that human isolates and animal isolates are largely distinct, particularly in areas with adequate on-farm biosafety and hygiene practices (Miltgen et al., 2022;Ludden et al., 2019;Day et al., 2019).Further, while clonal diversity was high among ESBL-EC linked to city areas, specific E. coli SNP clusters appeared to be dominant among ESBL-EC from chicken farms and pig farms.The increased prevalence of specific lineages such as E. coli ST155 among isolates from chicken farms, and ST44, ST1081 and ST14801 linked to pig farms may reflect their adaptation to the host animals on those farms.Finally, several STs that were identified less frequently in this study have also been described in E. coli from urban rats elsewhere, including rodent E. coli ST34 from Hong Kong and from Vienna (Desvars-Larrive et al., 2019;Ho et al., 2015), ST117 from Guadeloupe (Guyomard-Rabenirina et al., 2020), ST156 and ST226 from Hong Kong (Ho et al., 2015), and ST1286 from Berlin (Guenther et al., 2012).However, these STs do not appear to be rat-specific, since they have been documented in E. coli from various sources according to the public database https://enterobase.warwick.ac.uk/(last accessed February 1, 2024).With the exception of E. coli ST744 (HK155) and E. coli ST648 (HK71), no potential international linkages at the HC10 level were detected.This contrasts with recent analyses of AMR E. coli retrieved from other sentinel species (predominantly migratory birds) which act as carriers and long-distance spreaders of international AMR E. coli clones from animals and humans across different environments (Brendecke et al., 2022;Medvecky et al., 2022;Wyrsch et al., 2022).
This study has some limitations.First, there is the restricted exploration of any putative plasmids harbouring the bla ESBL genes described in this study, precluding a better estimation of the role of potential horizontal gene transfer between E. coli carried by rats from different locations.Complementary investigations are therefore needed to clarify the characteristics of bla ESBL -carrying plasmids from E. coli from this study.Second, this study is focused exclusively on the model organism E. coli and resistance to third generation cephalosporins.Therefore, the results cannot be generalized to other Enterobacterales or to other genetic resistance determinants that may be carried by the rats investigated this study.It is possible that the inclusion of other species and other AMR genes may have led to further conclusions.
Rats are considered good sentinel animals to detect discrete variations of AMR within their habitat, because they forage and feed on organic matter such as human food waste, domestic garbage, wastewater, or livestock food sources (de Cock et al., 2023;Strand and Lundkvist, 2019).However, the distribution of phylogenetic groups among the E. coli in this study indicate that host-specificity of the rat gut may compromise the accuracy of reflecting E. coli within the environment with regard to E. coli B2 and D. Nevertheless, our data strongly suggest that rats are colonized with human or livestock adapted ESBL-EC and may contribute to the dissemination of ESBL-EC within city areas and within farm environments, respectively, with implications for public health and animal welfare.Our study points to the need to control the occurrence of rats inhabiting city areas and farms, for example by the application of sustainable pest management strategies, to reduce the spread of AMR in the environment.

Conclusions
The prevalence of MDR ESBL-EC among peridomestic rats in city areas and animal farms in Hong Kong is high, highlighting the potential role of rats to disseminate AMR in the environment.The abundance of AMR genes that confer resistance to multiple antimicrobials identified among the rodent ESBL-EC suggests strong selective pressures within the environments inhabited by peridomestic rats.The populations of rodent ESBL-EC from city areas, chicken farms and pig farms were genetically different, indicating a certain degree of partitioning between the human and animal locations.Efficient pest control in city areas and on livestock farms may mitigate the risk of dissemination of MDR ExPEC to humans, and of animal adapted ESBL-EC to livestock respectively, to safeguard public and animal health and to reduce the impact of AMR on the environment.This study contributes to current understanding of the spatial distribution and variability of ESBL-EC occurring in rats in ecologically diverse locations.
Notably, the focus on E. coli may underestimate the prevalence of other AMR pathogens carried by rats.Although the identification of plasmid sequences suggests the potential for horizontal transfer events which may influence the characteristics of rat-borne AMR pathogens, this conclusion is limited without additional analysis.Future research should therefore include a wide range of rat populations in different environments, cover multiple bacterial species, and examine the impact of horizontal gene transfer.

Fig. 3 .
Fig. 3. Shared extended-spectrum ß-lactamase (bla ESBL ) genes identified in 134 ESBL-producing E. coli (ESBL-EC) isolated from 130 rats trapped in city areas, chicken farms, and pig farms in Hong Kong.The intersection areas of the circles represent the shared bla ESBL variants between two or three locations.Locations that did not share identical bla ESBL variants are indicated with ø.The number of isolates harbouring a bla CTX-M variant is indicated in brackets.A total of 18 ESBL-EC contained more than one bla ESBL .Only locations that contained two or more ESBL-EC are shown.For simplification, HK182 (bla CTX-M-15 ) from a horse-riding school is not listed (for details see main text).

Fig. 4 .
Fig. 4. Shared sequence types (STs) of 134 extended-spectrum ß-lactamase-producing E. coli (ESBL-EC) isolated from 130 rats trapped in city areas, chicken farms, and pig farms in Hong Kong.The intersection areas of the circles represent the shared ST between two or three locations.Locations that did not share any ESBL-EC STs are indicated with ø.The number of isolates belonging to an ST is indicated in brackets.Where no bracketed numbers are indicated, the number was one isolate.A total of 18 ESBL-EC contained more than one bla ESBL .Only locations that contained two or more ESBL-EC are shown.For simplification, HK182 (ST4681) from a horse-riding school is not listed (for details see main text).

Fig. 5 .
Fig.5.Maximum-likelihood phylogenetic tree of 134 extended-spectrum ß-lactamase-producing E. coli (ESBL-EC) isolated from rats in Hong Kong.Each isolate is annotated with its sequence type (ST), bla ESBL variant(s), host source (Rattus spp.), and rat trapping location in Hong Kong.Putative transmission clusters (<10 cgSNP pairwise distance) are shaded in pale blue; cluster numbering is indicated within the shaded area (for details see main text).The tree is based on 97,905 variant sites identified in a 1.3 Mbp core genome alignment.The scale bar indicates the number of substitutions per core genome alignment site.The tree was visualized using iTOL(Letunic and Bork, 2021).