First Detection of International High-Risk bla KPC-2 -Harbouring Escherichia coli Pandemic Lineage ST648 in Pet Food Packages

The continued worldwide increase in pet ownership has signi ﬁ cantly boosted the growth of the pet food industry accompanied by new food safety risks and challenges. This study was designed to determine the occurrence and molecularly characterize multidrug-resistant (MDR) Enterobacterales in pet food. Eighty-six (86) packages of dry and wet pet food purchased in different retail stores were screened for carbapenem-resistant Enterobacterales (CRE). Antimicrobial susceptibility testing was performed by agar dilution technique using EUCAST/BrCAST recommendations. Blue-Carba test was further used to screen for carbapenemase-producing isolates. Isolated CRE strains were identi ﬁ edat the species level using matrix-assisted laser desorption/ionisation time-of-ﬂ ight mass spectrometry(MALDI-TOF MS). Detection of carbapenemase-encoding genes was carried out by PCR, Sanger sequencing, and whole genome sequencing (WGS). A total of 15 (17.4 % ) MDR-CRE ( Escherichia coli ( n = 2), Enterobacter cloacae ( n = 10), Leclercia adecarboxylata ( n = 2), and Cronobacter spp. ( n = 1)) were isolated from 86 pet food samples. In addition to being resistant to beta-lactams, the Gram-negative bacterial isolates were also resistant to aminoglycosides, ﬂ uoroquinolones, and tigecycline. Interestingly, two carbapenem-resistant E. coli isolates harboured bla KPC-2 gene. WGS analysis of the two bla KPC-2 -producing E. coli isolates revealed that they both belong to ST648 and serotype O153:H2 group. The genetic context of the bla KPC-2 showed that they were carried by an IncN plasmid on a Tn 4401b transposon element. To the best of our knowledge, this is the ﬁ rst description of bla KPC-2 -harbouring E. coli ST648 pathogens in pet food. The detection of bla KPC-2 -harbouring E. coli ST648 pandemic high-risk lineage in pet food is worrisome and a serious “ One Health ” issue. Therefore, pet food should be considered as a potential vehicle for the transmission of MDR pathogens to companion animals, and a risk factor for the dissemination of these bacterial pathogens to pet animals and their human guardians.


Introduction
The increasing global spread of antimicrobial resistance (AMR) is a critical issue that is no longer restricted to hospital settings, but also represents an important and underestimated growing problem involving food safety [1].In the last decade, the frequency of pet ownership, especially dogs and cats, has significantly increased worldwide.This worldwide increase in pet ownership in recent years, especially in the United States with the largest pet population (223 million), followed by China (190 million), and Brazil (139.3 million) (http://petbrasil.org.br/importer), has been linked to an exponential growth of the pet food industry.This has been accompanied by new food safety risks such as AMR and pet food contamination by bacterial pathogens [2][3][4].
The worldwide spread of carbapenem-resistant Enterobacterales (CRE), including bla KPC-2 -harbouring bacterial pathogens, is a serious concern and a challenging public health problem worldwide [5,6].Infections caused by CRE are associated with high morbidity and mortality rates due to the limited availability of therapeutic alternatives and the lackadaisical development of new antimicrobials [6].The World Health Organization (WHO) considers CRE as global priority pathogens in critical need of next-generation antimicrobials and new control strategies [7,8].The five major carbapenemases identified in Gram-negative bacterial pathogens are those belonging to the groups KPC, NDM, VIM, IMP, and OXA-48 [9].KPC-2 are endemic in low-and middle-income countries, especially in Latin America (such as Colombia, Brazil, Chile, Bolivia, and Argentina) and in some developed countries (such as USA and China) [10,11].
Reports on pet food contamination by bacterial pathogens, especially CRE strains are very scarce as little or no attention has been paid to this surveillance in many countries and regions of the world, especially in Latin America, including Brazil.However, despite this neglect, Seiffert et al. [12] reported the presence of bla OXA-48 (13.3%) gene and other β-lactamase-encoding genes such as bla CTX-M-15 (53.3%), bla CMY-4 (20%), and bla VEB-4-like (6.7%) in 30 packages of pet food samples obtained from three different retail outlets in Europe.The contamination of pet food by important foodborne pathogens has also been reported in a study carried out in the United States where 16.3%, 7.6%, and 4.1% of raw pet foods ordered online were positive for Listeria monocytogenes, Salmonella spp., and Escherichia coli, respectively [13].
Unlike the European Union and other countries that have implemented strict safety guidelines for pet food, e.g., the European Union Regulation (EC) No 142/2011 [14] and the USA Compliance Policy Guide Sec 690.800 in Food for Animals [15], Brazil and some other countries in the world are yet to establish strong guidelines on pet food safety standards.The established health safety guidelines in Brazil majorly focus on food-producing animals (livestock) and their products, and human health.In Brazil, the "Ministry of Agriculture and Livestock" oversees the inspection and supervision of products for animal feed, approval of technical regulations on hygienic-sanitary conditions and good manufacturing practices (GMPs) for establishments that manufacture animal feed or products intended for animal feed, and registration of pet food production companies; however, there are no well-defined standards/guidelines on pet (dogs and cats) food safety.The unavailability of welldefined guidelines on pet food safety standards in Brazil and other countries could contribute to the spread of antimicrobial resistance with significant public health impact, especially if contaminated pet food products are consumed by companion animals which are in close proximity with their human guardians.
The continued increase in AMR in pets, humans, and food-producing animals with their products has emerged as a leading public health threat in the 21st century [4].Investigating the contamination of pet food by bacterial pathogens would be a major step in understanding the genesis of the zoonotic transmission of bacterial pathogens from companion animals to humans.
Herein, this study aimed to assess pet food as potential sources of carbapenemase-producing bacterial pathogens in the city of Sao Paulo, Brazil.To the best our knowledge, this study is the first to report the detection of bla KPC-2 -harbouring E. coli ST648 in pet food in Brazil.

Materials and Methods
2.1.Sample Collection.Eighty-six (86) pet food package samples (dry and wet) belonging to 12 different brands were purchased from different retail stores in São Paulo, Brazil, between February and June 2022.All the pet food samples were manufactured in Brazil.The indicated ingredients on the packages of the pet food were chiefly byproducts of poultry (chicken and turkey viscera, chicken and turkey meat), other food-producing animals (such as beef, pork meat, lamb meat, animal liver, animal bones), fish, egg, and some plantbased food sources (such as rice, corn, wheat, oats, and sorghum; Table 1).Collected samples were delivered to the laboratory on the same day of collection within 2 hr for bacteriological analysis.
2.2.Sample Processing.Exactly 25 g of pet food samples were enriched in 225 mL of Luria-Bertani (LB) broth and homogenized for 5 min before incubation at 37°C for 24 hr.After incubation, 100 µL of the broth was transferred into different microtubes which contained 900 µL of tryptic soy broth (TSB) and different concentrations of meropenem (MEM) and vancomycin (VAN): Tube 1 (2 µg/mL of MEM + 4 µg/ mL of VAN), and Tube 2 (2 µg/mL of MEM).Controls were also included.Inoculated microtubes were then incubated under selective pressure at 37°C for 18-24 hr.After incubation, 10 µL of broth from tubes showing turbidity (growth) were inoculated onto MacConkey agar (Oxoid, UK) and incubated at 37°C for 24 hr.Recovered pure colonies were further identified at the species level by Matrix-Assisted Laser Desorption Ionization-Time-of-Flight Mass Spectrometry (MALDI-TOF MS) using Microflex LT spectrometer and Biotyper™ 3.3 software package (Bruker Daltonics™, MA, USA), according to the manufacturer's instructions.Identified bacterial colonies were preserved at −80°C in trypticase soy broth (TSB) with 15% glycerol for further analysis.

2
Transboundary and Emerging Diseases 2.4.Phenotypic Detection of Carbapenemase Production.All carbapenem-resistant isolates were further screened for carbapenemase production using the Blue-Carba test, as previously described [18].

Clonal Relationship Determination by Enterobacterial
Repetitive Intergenic Consensus (ERIC) PCR.The clonal relationship was initially investigated by ERIC-PCR, as previously described [22].The amplified PCR products were electrophoresed in 1.5% agarose gel for 2 hr and visualized in the gel documentation system.The DNA band patterns were visually observed and analyzed.

Whole Genome Sequencing (WGS) and Genomic Data
Analysis.Genomic DNA was extracted from bacterial pellets with ZymoBIOMICS DNA miniprep kit (Zymo Research) following manufacturer's instructions.The DNA purity and quantity were assessed using the Nanovue Plus (GE Healthcare) and a Qubit double-stranded DNA (dsDNA) broad range (BR) assay kit (Fisher Scientific).The sequencing libraries were prepared with the rapid barcoding sequencing kit 96 (SQK-RBK110.96;Oxford Nanopore Technologies [ONT]).Barcoded samples were loaded in a R.9.4.1 flow cell (FLO-MIN106; Oxford Nanopore Technologies Ltd.) and sequenced in a MinION Mk1C device (ONT).The sequencing data contained within the fast5 files were base-called and demultiplexed using the MinKNOW software equipped with the Dorado base-calling algorithm (version 7.0.8).This process was operated in "super-high accuracy mode" on the MinION sequencing device.Additionally, adapter sequences were trimmed, and any reads that did not meet predefined quality standards were excluded.

Transboundary and Emerging Diseases
The assembled whole genomes of the bla KPC-2 -harbouring E. coli isolates (M43 and M49) were annotated in BV-BRC (https://www.bv-brc.org/)platform using RAST tool kit (RASTtk).Based on the annotation statistics, the two sequenced genomes were of good quality and identified as E. coli.Each of the two assembled genomes had, respectively: 24 and 18 contigs; total length of 5,107,014 and 5,058,972 bp; and G + C content of 50.57% and 50.59%.
Genomic analysis revealed that both M43 and M49 E. coli isolates also carried tet(B) genes in their resistome.WGS analysis also identified the sitABCD system encoding resistance to heavy metals and mediating the transport of manganese and iron which may contribute to oxidative stress resistance.Additionally, chromosomal mutations in gyrA, gyrB, parC, parE, pmrA, pmrB, rrsB, rrsC, rrsH, rpoB, folP, and AmpC-promoter mediating AMR were also observed.The chromosomal mutations in gyrA, gyrB, parC, and parE might be responsible for the high MICs observed for the phenotypic resistance to fluoroquinolones while the chromosomal mutations in rrsB, rrsC, and rrsH be responsible for the MICs recorded for the resistance to aminoglycosides employed in this study.In silico analysis showed that the two E. coli isolates possessed several relevant virulence-encoding genes characteristic of extraintestinal pathogenic E. coli (ExPEC) such as: chuA (outer membrane hemin receptor), kpsMIII (ABC-type export system and group 3 capsule synthesis), fimH (fimbriae type I), sitA (iron transport protein), air (Enteroaggregative immunoglobulin repeat protein), eilA (transcriptional regulator), and lpfA (long polar fimbriae protein), along with an average of 484 pathogenic protein families (https://cge.food.dtu.dk/services/PathogenFinder/).Multilocus sequence typing (MLST) and serotype analyses revealed that both M43 and M49 E. coli isolates belong to ST648 (adk-92, fumC-4, gyrB-87, icd-96, mdh-70, purA-58, recA-2) and serotype O153:H2 group.Plasmidome analysis showed that both M43 and M49 harboured an IncN plasmid.Analysis of the IncN plasmids of M43 and M49 showed that they both harboured bla KPC-2 gene and also have a conserved backbone with a reference IncN plasmid (Accession number: AY046276).
The bla KPC-2 gene of M43 and M49 were both located on a Tn3-family transposon known as Tn4401b (Accession number of the reference genome: EU176014.1),and flanked upstream by ISKpn7, tnpA, tnpR, and downstream by ISKpn6 in an ∼50.3-kbIncN plasmid to form a tnpR-tnpA-ISKpn7-bla KPC-2 -ISKpn6 core structure (Figure 2).The insertion sequences ISKpn7 and ISKpn6 are well-recognized putative transposition helper which are additionally flanked upstream and downstream by plasmid conjugative transfer endonuclease proteins and series of hypothetical proteins.Accessory genes, tnpA (a transposase) and tnpR (a resolvase), have also been recognized to be involved in transposition immunity.

Discussion
CRE have been recognized as critical priority pathogens by the WHO (https://www.who.int).The continued emergence and spread of CRE is a growing public health challenge in veterinary and human medicine [24].Pet food might serve as potential sources and vehicles for the transmission of antimicrobial-resistant bacterial pathogens, including CRE to companion animals, and subsequently to humans (especially pet guardians) [25].The zoonotic transmission risk of CRE via pet food is noteworthy as those microorganisms are recognized globally as major threats to public health.In this study, 15 (17.4%)CRE were isolated from 86 pet food samples.These CRE isolates were also observed to be multidrugresistant (MDR) as they exhibited resistance to at least three different antimicrobial classes.Interestingly, two (13.3%)E. coli isolates from different batches of pet food harboured bla KPC-2 genes which were further confirmed by sequencing.ERIC-PCR results also showed that these two bla KPC-2 -harbouring E. coli isolates are clonally related.WGS results showed that the two bla KPC-2 -harbouring E. coli isolates belong to the high-risk epidemiological lineage ST648 and serotype O153:H2 group.The E. coli ST648 is a successful evolutionary lineage and has emerged as a pandemic clone, being globally reported in humans, companion animals, and the aquatic environment [26,27].Our study has been able to demonstrate that pet food is a possible route for CRE transmission which has been overlooked by the scientific community.However, more studies that would holistically evaluate the transmission of clinically important bacterial pathogens (such as CRE) from contaminated pet food to companion animals (especially dogs and cats), and pet guardians/owners would be very valuable in order to comprehensively understand the epidemiological gaps concerning the acquisition of CRE within the One Health concept.
The genetic context of the bla KPC-2 showed that they were carried on a Tn4401b transposon in an IncN plasmid.IncN plasmids harbouring bla KPC are known to be mobile genetic elements of worldwide epidemiological importance.Since at least 2015 to date, the circulation of bla KPC-2 -harbouring IncN plasmid has been reported among Enterobacterales infecting humans and animals in Brazil [28], Colombia [29], USA [30], and Germany [31].The epidemiological success of carbapenem-resistant bacterial pathogens has been attributed to plasmids carrying bla KPC -like genes that are associated with the Tn4401 transposon [32][33][34].Of interest is the IncN which has been reported to be highly disseminated, especially among international clones of E. coli (ST131 and ST648) [32,33].
The prevalence of bacterial pathogens in pet food has been investigated in the United States [13] where 4.1% out of 196 frozen raw pet food samples ordered online were positive for E. coli.Nüesch-Inderbinen et al. [35] also reported the isolation of MDR E. coli and other members of the Enterobacterales in pet food in Switzerland.Interestingly, we also isolated two E. coli isolates from pet food package samples in this study.Raw pet food samples contaminated by E. coli were also identified as incriminating factors in a series of investigations on foodborne illnesses associated with E. coli in companion animals [35,36].A previous study in Europe (Netherlands) reported that raw pet food is more contaminated with ESBL/ AmpC-producing Enterobacteriaceae than non-raw pet food products [37].Additionally, Baede et al. [37] observed that raw pet food is an important risk factor for ESBL/AmpC shedding in household cats, and that the consistent exposure of household cats to raw pet food products appears to be linked to shedding of ESBL-positive Enterobacteriaceae, rather than mere gut colonization.
Even though there are pockets of reports on the colonization of companion animals by MDR-GNB [2,[38][39][40], studies on bacterial contamination of pet food are very scarce as proper attention has not been directed to this investigation.Even in Latin America, including Brazil, there are no surveillance studies on pet food contamination by bacterial pathogens, to the best of our knowledge.
In the last two decades, plasmid-mediated β-lactamaseencoding genes in Enterobacterales have been widely reported in Brazil, with bla KPC-2 -producing Klebsiella pneumoniae being the most endemic [10,42].Even though the pathogenesis of MDR-GNB harbouring different types of carbapenemaseencoding genes such as bla KPC -like, bla NDM -like, and bla OXA-48like have been widely reported in humans, food-producing animals (livestock), and companion animals [2,6,[38][39][40]42]; information on pet food contamination by these MDR bacterial pathogens, especially those harbouring bla KPC-2 , is very scarce.Seiffert et al. [12] reported bla OXA-48 , which encodes for a class D carbapenemase in 30 analyzed pet food samples.Contrastingly, the detection of carbapenem-resistant pathogens harbouring only bla KPC-2 gene, without other types of carbapenemaseencoding genes in our study further depicts and confirms its regional dominance in Latin America, especially in Brazil.
It is interesting to know that raw materials/ingredients used in the manufacture of the pet food in our study chiefly contained byproducts of poultry (chicken and Turkey viscera, chicken and Turkey meat), other food-producing animals, fish, egg, and some plant-based food sources, according to the stated compositions on their packages.These byproducts from food-producing animals have been reported to harbor antimicrobial-resistant pathogens, including CRE [3,9,24,41].However, considering that packaged pet food samples go through sterilization techniques such as heat treatment and irradiation, it is still not clear how they were contaminated by bacterial pathogens, especially bla KPC-2 -harbouring E. coli ST648 as reported in our study.A limitation of our study was our inability to completely ascertain the source of the pet food contamination by CRE as we did not have access to the pet food ingredients/constituents nor the production machinery.However, we suspect that contamination of the pet food might possibly be from contaminated food ingredients, machinery components/equipment, environmental contamination, or any of the production processing steps, including the final packaging of the pet food.
Reports on pet food contamination by bacterial pathogens have led to arrays of pet food recalls in the United States 6 Transboundary and Emerging Diseases by the FDA [15,43].The consumption of these CREcontaminated pet food by companion animals might negatively impact companion animal health and further exacerbate the zoonotic transfer potentials of these CRE from companion animals to humans [9].Data from our study will create awareness on the significance of identifying the sources of pet food contamination as it might be an important source of clinically relevant MDR bacteria and their associated pathogenic factors.
Hence, more studies are needed to elucidate the possible sources of pet food contamination.

Conclusions
In conclusion, we report for the first time in pet food, the detection of an international high-risk bla KPC-2 -harbouring E. coli pandemic lineage ST648 in South America.The bla KPC-2 gene was inserted in the Tn4401b transposon carried by an IncN plasmid.The detection of CRE in packaged pet food is a critical public health problem with great significance to "One-Health" as they could serve as potential sources for the spread of MDR pathogens to companion animals, and possibly to humans through zoonotic transmission events.It is therefore imperative for the government and stakeholders in the pet food industry to reevaluate their production processes and techniques, especially proper sterilization of pet food ingredients and ensuring good Hazard Analysis Critical Control Point (HACCP) manufacturing practices to completely eradicate MDR bacterial pathogens from pet food products.This will strongly help to curb the dissemination of clinically important CRE in human and veterinary medicine.

Table 1
Carbapenem-Resistant Bacterial Pathogens and Genetic Characterization.Only two carbapenem-resistant E. coli isolates (M43 and M49) were positive for Blue-Carba test (Table2).Interestingly, the two (13.3%)carbapenemresistantE. coli isolates from two different batches of brand "F" pet food sample (Table1) harboured bla KPC-2 gene, a class A carbapenemase-encoding gene (Table2) which was further confirmed by sequencing.The DNA sequences of the amplified bla KPC -like genes harboured by the E. coli isolates in our study revealed that they were 100% identical to the reference nucleotide sequence of bla KPC-2 gene (NCBI Reference Sequence: ON412784.1)deposited at the GenBank database/ National Center for Biotechnology Information (NCBI) server (https://blast.ncbi.nlm.nih.gov/Blast.cgi).Interestingly, ERIC-PCR results showed that the two bla KPC-2 -harbouring E. coli isolates belong to the same clone.These results were confirmed by WGS.However, all E. coli isolates were negative for bla NDM-1 , bla OXA-48 , bla OXA-10 , and bla OXA198 carbapenemase-encoding genes.

TABLE 2 :
Antimicrobial MICs for GNB isolated from pet food samples.