One health clones of multidrug-resistant Escherichia coli carried by synanthropic animals in Brazil

WHO priority pathogens have disseminated beyond hospital settings and are now being detected in urban and wild animals worldwide. In this regard, synanthropic animals such as urban pigeons (Columba livia) and rodents (Rattus rattus, Rattus norvegicus and Mus musculus) are of interest to public health due to their role as reservoirs of pathogens that can cause severe diseases. These animals usually live in highly contaminated environments and have frequent interactions with humans, domestic animals, and food chain, becoming sentinels of anthropogenic activities. In this study, we report genomic data of Escherichia coli strains selected for ceftriaxone and ciprofloxacin resistance, isolated from pigeons and black rats. Genomic analysis revealed the occurrence of international clones belonging to ST10, ST155, ST224 and ST457, carrying a broad resistome to beta-lactams, aminoglycosides, trimethoprim/sulfamethoxazole, fluoroquinolones, tetracyclines and/or phenicols. SNP-based phylogenomic investigation confirmed clonal relatedness with high-risk lineages circulating at the human-animal-environmental interface globally. Our results confirm the dissemination of WHO priority CTX-M-positive E. coli in urban rodents and pigeons in Brazil, highlighting potential of these animals as infection sources and hotspot for dissemination of clinically relevant pathogens and their resistance genes, which is a critical issue within a One Health perspective.


Introduction
WHO critical priority pathogens, which include ESBL-producing Enterobacterales, have broken the boundaries of hospital settings and their prevalence in both hospital-and community-acquired infections is on a rise [1,2,3]. Moreover, ESBL-producing Escherichia coli has been lately reported in wild animals, including urban wildlife [1,4,5,6]. In this regard, CTX-M positive E. coli has become a threat to global health and a One Health challenge, due to its resistance to several commonly used antibiotics, its ability to colonize or infect different hosts either through contact or through food/water contamination, presence in different environments, rapid spread, and worldwide distribution [1]. Synanthropic animals, such as urban pigeons (Columba livia) and rodents (Rattus norvegicus, Rattus rattus and Mus musculus) are of interest in public health due to their frequent interactions with humans, animals and food chains, their habits of foraging in human leavings and garbage and sheltering in highly contaminated environments, such as sewers and polluted river banks (in the case of rats), and their role as reservoirs of zoonotic agents that can cause severe human and animal disease [7,8,9]. Such characteristics also make synanthropic animals sentinels of anthropogenic activities, being potential indicators of clinically relevant drug-resistant Enterobacterales, such as CTX-positive E. coli, beyond hospital settings ( [10]; [11][12][13][14][15][16]; [17]; [18]). In this study, we report genomic data and phylogenomic analysis of multidrug-resistant E. coli, including four CTX-M-positive isolates from feral pigeons and black rats.

Field capture and sample collection
Between September 2018 and November 2019, biologists of the Zoonoses Surveillance Division of São Paulo city, Brazil, captured 22 feral urban rodents (1 R. norvegicus and 21 R. rattus), and 22 diseased (n = 15) or dead (n = 7) feral urban pigeons (C. livia) within the Division's headquarters and surrounding areas. Rodents were captured using cage traps in rodent control activities. By the time of collection, 16 rodents were living and 6 were dead. Rectal and oral samples were collected, as well as feces deposited in the trap (when available). Species identification of the rats were made by rodent specialist's biologists working in the Nucleus of Surveillance and Control of Synantropic Animals at the Center for Zoonoses Control, in São Paulo; using guidelines provided by the Brazilian Ministry of Health (https://bvsms.saude.gov.br/bvs/pub licacoes/manual_roedores1.pdf). In this regard, R. norvegicus and R. rattus were captured in an urban area (Table 1) near to drainage canals and the Tietê River, which is affected by the direct discharge of anthropogenic pollutants (including domestic sewage and hospital wastewater) in São Paulo, the largest and most populous city in Brazil. Pigeons were collected during the investigation of a Newcastle Disease Virus (NDV) outbreak [19]. Cloacal and ingluvial samples were collected from 15 diseased and 7 dead pigeons, using transport swab with Amies and charcoal medium (Copan Italia, Brescia, Italy).

Whole genome sequencing
Genomic DNAs from all ESBL-producing E. coli strains, isolated from pigeons (P2C1, P22C2 and P32P1) and rodent (R6R1), and from the ciprofloxacin-resistant E. coli strain R7F1 isolated from rodent were

Results and discussion
Four E. coli strains from 4 different rodents were isolated using drugsupplemented agar. In this regard, while 2 E. coli were recovered from ceftriaxone-supplemented agar, another 2 isolates were recovered from ciprofloxacin-supplemented agar. From pigeon samples, 8 E. coli strains were isolated from 6 different birds, of which 3 isolates have grown on both drug-supplemented plates. Additionally, 3 E. coli isolates have grown only on ceftriaxone-containing agar plates, and other 2 isolates were isolated from ciprofloxacin-containing agar plates. Antimicrobial susceptibility tests were performed for all 12 isolates (Fig. 1).
Higher rates of antimicrobial resistance were observed in pigeon isolates than in rodent isolates (Fig. 1). Moreover, ESBL-producing isolates were observed in 9.1% of rodents and 13.1% of pigeons. Higher resistance rates and prevalence in pigeons is possibly due to the fact that pigeons collected for this study were captured in an NDV outbreak context, and all of them were diseased or dead, thus probably more susceptible to opportunist pathogens.
Isolates displaying resistance to 3 or more classes of antibiotics were considered multi-drug resistant (MDR) [23]. For WGS analysis 5 E. coli isolates were selected based on a MDR profile for broad-spectrum cephalosporins, ciprofloxacin, co-trimoxazole (trimethoprim-sulfamethoxazole), tetracycline or gentamicin ( Fig. 1; Table 1): R6R1, a rectal isolate from a black rat captured dead; and R7F1, isolated from rat feces; P2C1, a cloacal isolate from a diseased NDV-positive pigeon; P22C2, a cloacal isolate from a non-tested diseased pigeon; and P32P1, an ingluvial isolate from a diseased NDV-negative pigeon.
In this study, the occurrence of ST10, ST155, ST224 and ST457 was confirmed in urban rats and pigeons (Table 1). In this respect, ST10 is a global E. coli lineage, found in different hosts, associated with extraintestinal pathogenicity that may carry different types of plasmids containing resistance genes [24,25]. In Brazil, high-risk clones of E. coli belonging to ST10 has been mostly identified in environmental and animal sources, including coastal waters, penguins, and cattle [26,27]. ST155 has been recently described as a potential food-borne pathogen, since it can cause disease in both poultry and humans [28,29,30]; having been recently identified in Brazil in humans, urban-impacted coastal waters and river fishes [31,26,32].
ST224 is a global clone frequently described in food production animals and retail food [33,34], as well as wild animals [35,36], having also been reported causing urinary tract infection in humans [37]. In Brazil, ST224 was identified in swine, companion animals and river fishes [32,34,38]. The ST457 has been recently described as an emerging extraintestinal pathogenic E. coli found mainly in wildlife and in food-producing animals [39], having also been reported in marine environment in Brazil [40], whereas transmission from companion animals to humans has been documented in China [41]. Moreover, this lineage has been associated to bloodstream infections [42].
On phylogenomic analysis, ST10 strains formed two clades, one for each fimH type, 30 and 54 (Fig. 2). Percentage of reference genome covered by all isolates was 76.97%, corresponding to 3,612,870 positions found in all analyzed genomes, and SNP counts ranged from 0 to 10,152 (Table S2).
On fimH54 clade, which includes R7F1 strain, SNP differences ranged from 0 to 3698, and half of the isolates had no resistance gene identified (Fig. 2). R7F1 strain is genetically close (186 SNPs) to a MCR-1-producing wild bird isolate from Spain, but resistomes and plasmidomes are different, e.g., while the isolate from Spain has no resistance gene for fluoroquinolones, R7F1 has the qnrB19 gene and mutations in QRDR, which highlights the importance of both horizontal transfer of resistance genes in mobile genetic elements and vertical transfer of mutations in resistance-determining regions for antimicrobial resistance spreading. On the other hand, on ST10fimH30 clade, which includes P22C2 strain, SNP counts ranged between 1 and 3678 (Table S2). Most of the isolates within fimH30 clade presented resistance genes for aminoglycosides, beta-lactams, fluoroquinolones, sulfonamides, trimethoprim, tetracyclines and/or hydrogen peroxide, as well as mutations in QRDR. P22C2 grouped with strains from human, companion animal, wild bird and environmental sources, mostly CTX-M-1 or CTX-M-8 producers, from Germany, France, Brazil, and Australia, with 57 and 185 SNP differences (Table S2).
Phylogenomic analysis of ST155 fimH121 revealed high diversity among strains, with 0 to 3047 SNP differences (Table S3). Percentage of reference genome covered by all genomes was 87.78%, corresponding to 4,254,315 positions found in all genomes. R6R1 grouped with relatively distant strains: 5 livestock isolates from China, next a human isolate from Tanzania. SNP count among strains within this clade ranged from 0 to 2247 (Table S3). Strikingly, while R6R1 presents resistance genes for aminoglycosides, beta-lactams, sulfonamides, trimethoprim and tetracyclines, we did not identify any resistance gene in the other strains within the clade (Fig. 3). The presence of Inc-type plasmids was only observed in R6R1, highlighting the importance of horizontal transfer.
Within ST224 fimH61 strains, 90.99% of reference genome was covered by all isolates, corresponding to 4,405,087 positions. SNP differences ranged between 0 and 1288 (Table S4). Phylogenomic analysis revealed that P32P1 is genetically very close to 4 strains from Switzerland, isolated from companion animal, human and environmental sources, sharing the same resistome and plasmidome (Fig. 4). Within this clade, SNP counts ranged from 3 to 16 (Table S4), which suggest this lineage is able to clonally spread to different hosts, highlighting a zoonotic/zooanthroponotic potential and the possible role of pigeons as reservoirs.
Finally, on phylogenomic analysis of ST457 fimH145, 91.46% of the reference genome was covered by all isolates, corresponding to 4,480,690 positions. SNP differences ranged between 1 and 2427 (Table S5). P2C1 grouped with 2 poultry isolates from United States, with SNP counts of 250 and 265 (Table S5). While P2C1 presents resistance genes for aminoglycosides, beta-lactams, sulfonamides, in the 2 closest isolates we only identified the sitABCD locus, related to hydrogen peroxide tolerance (Fig. 5).
Plasmid analysis of phylogenetically related E. coli ST10 and ST254 showed that IncQ1 plasmids are widespread in those strains, and usually carries aminoglycosides and sulfonamides resistance genes. However, an IncQ1 plasmid from ESC_LB0138AA additionally carried a chloramphenicol/florfenicol efflux MFS transporter floR gene and thus regarded as multidrug-resistance (MDR) plasmid. On the other hand, our pigeon strain P22C2 harbored a MDR IncQ1 plasmid carrying macrolide 2 ′phosphotransferase II [mph(B)], a class I integron containing drfA1-sul1-qacEΔ1-sul2 genetic array, and the mercury resistance operon (mer-EDACPTR) (Fig. 6).
Due to limitations of short-read sequencing technology, it was not possible to obtain complete nucleotide sequences of the plasmids, but further analysis revealed that IncQ1 plasmids were found to carry antimicrobial, disinfectants, and heavy metal resistance genes in P22C2 and P32P1 (Fig. 6). The ESBL CTX-M-types identified in this study were in contigs of short length, and thus its precise location could not be determined. However, in P22C2 and P2C1, bla CTX-M-2 and bla CTX-M-8 genes were mediated by IS91 family transposase, whereas in R6R1, bla CTX-M-8 was identified upstream of a IS4 insertion sequence. The bla CTX-M-1 gene was located 46 bp downstream of a cupin fold metalloprotein (wbuC) in P32P1 strain (Fig. S1).
Although this study has limitations such as convenience sampling and a small number of animals collected in just one place, the detection Fig. 1. Results of antimicrobial susceptibility testing for each Escherichia coli isolate from pigeons and rodents obtained in ceftriaxone-and ciprofloxacinsupplemented agar, as well as resistance rates for each tested antibiotic. Fig. 2. In A, phylogenetic tree plotted in a 180 • arc, with 62 Escherichia coli ST10 strains, their isolation source, presence/absence of antimicrobial resistance genes for different antimicrobial classes, and country of collection. In B, highlighted cluster in "A" with fimH54 strains, their source of isolation, resistome, plasmidome, country and year of collection. In C, highlighted cluster in "A" with fimH30 strains, their source of isolation, resistome, plasmidome, country and year of collection.
of ESBL-producing Enterobacterales in such animals is of concern, due to the proximity of synanthropic animals and human populations in urban environment. However, with this dataset it is not possible to neither compare different regions, nor analyze correlations with socioeconomic features. Therefore, studies with probabilistic sampling are needed in order to assess the role urban wildlife play in antimicrobial resistance spreading and transmission. Animals and carcasses collected in pest control activities may be valuable to elucidate how these animals may acquire, carry, spread and transmit antimicrobial resistant microorganisms into their life range.
In summary, synanthropic animals such as urban rats and pigeons are frequently exposed to contaminated places they use for shelter or foraging. Their behavior favors frequent interactions with humans, animals (companion, food-producing and wild), food chains, urban waterways, sewage, and domestic garbage, making them potential reservoirs and transmitters of antimicrobial resistant pathogens. Despite limitations in this study, phylogenomic results corroborates this hypothesis, once the isolates from rats and pigeons are closely related to isolates from different hosts and sources, such as humans, companion animals, environments, livestock, poultry, and other wild animals, suggesting that these animals may contribute with dissemination of WHO priority pathogens through the human-animal-environment interface, which is a critical issue within a One Health perspective.

Author statement
We declare that the manuscript "One Health clones of multidrugresistant Escherichia coli carried by synanthropic animals in Brazil" by Elder Sano, Fernanda Esposito, Herrison Fontana, Bruna Fuga, Adriana Cardenas-Arias, Quézia Moura, Brenda Cardoso, Gladyston C. V. Costa, Tatiana C. M. Bosqueiro, Juliana A. Sinhorini, Eduardo de Masi, Caroline C. Aires, and Nilton Lincopan, has not been published before and is not under consideration for publication elsewhere.
All authors made relevant contributions to the development of the research, the manuscript has been read and approved by all named authors and confirm that the order of authors listed in the manuscript has been approved by all of us. We also affirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome.
We understand that the Corresponding Author is the sole contact for the Editorial process. He is responsible for communicating with the other authors about progress, submissions of revisions and final approval of proofs. We confirm that we have provided a current, correct email address.
We are providing details in the table below of each author contribution to the submitted manuscript: Fig. 5. In A, phylogenetic tree plotted in a 180 • arc, with 31 E. coli ST457 fimH145 strains, their isolation source, presence/absence of antimicrobial resistance genes for different antimicrobial classes, and country of collection. In B, highlighted cluster in "A", with source of isolation, resistome, plasmidome, country and year of collection for each strain.

Declaration of Competing Interest
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.

Data availability
Genomic data ae publicly available on NCBI Fig. 6. Schematic representation of IncQ1 plasmids identified in E. coli ST10 and ST224. Arrows represents coding sequences and are labelled by color as follows: purple, antimicrobial resistance genes; red, disinfectants compounds or heavy metal resistance; shadow blue, plasmid replication proteins; light blue, mobile genetic elements; gray, hypothetical proteins or other genes. Dark gray shadows represent regions of homology. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)