Association between the blaCTX-M-14-harboring Escherichia coli Isolated from Weasels and Domestic Animals Reared on a University Campus

Antimicrobial-resistant (AMR) bacteria affect human and animal health worldwide. Here, CTX-M-14-producing Escherichia coli isolates were isolated from Siberian weasels (Mustela sibirica) that were captured on a veterinary campus. To clarify the source of bacteria in the weasels, we examined the domestic animals reared in seven facilities on the campus. Extended-spectrum β-lactamase (ESBL)-producing E. coli were isolated on deoxycholate hydrogen sulfide lactose agar, containing cephalexin (50 μg/mL) or cefotaxime (2 μg/mL), and were characterized with antimicrobial susceptibility testing, pulsed-field gel electrophoresis (PFGE), replicon typing, and β-lactamase typing analyses. Next-generation sequencing of the ESBL-encoding plasmids was also performed. CTX-M-14 producers isolated from both domestic animals and weasels were classified into six clusters with seven PFGE profiles. The PFGE and antimicrobial resistance profiles were characterized by the animal facility. All CTX-M-14 plasmids belonged to the IncI1 type with a similar size (98.9–99.3 kb), except for one plasmid that was 105.5 kb in length. The unweighted pair group method with arithmetic mean (UPGMA) revealed that the CTX-M-14 plasmid in the weasel isolates might have the same origin as the CTX-M-14 plasmid in the domestic animals. Our findings shed further light on the association of antimicrobial resistance between wild and domestic animals.


Introduction
The distribution of antimicrobial-resistant (AMR) bacteria is a significant concern worldwide. AMR bacteria in humans are linked to antimicrobial resistance in farm and wild animals [1]. Some wild animals are frequently found in or around human societies and domestic animal facilities [2][3][4] and can transfer AMR bacteria to humans and domestic animals within their environment. Additionally, horizontal gene transfer between bacteria via plasmids can potentially lead to the dissemination and proliferation of AMR bacteria [5]. The transfer of plasmids among several bacterial isolates diversifies the distribution of AMR bacteria.
Bacteria that produce CTX-M-type extended-spectrum β-lactamases (ESBLs) are resistant to third and fourth generation cephalosporins, which are commonly used in modern and veterinary medicine. CTX-M-producing bacteria are distributed worldwide and have been reported in wild animals [2,4,5]. In 2015, we isolated CTX-M-14-producing Escherichia coli from wild Siberian weasels (Mustela sibirica) present at the Gifu University campus (Gifu Prefecture, Japan). Some studies in Japan have reported the occurrence of several bla CTX-M genes in E. coli isolated from the feces of domestic animals, such as bla CTX-M-2 , bla CTX-M-14 , and bla CTX-M-25 in chicken isolates [6], bla CTX-M- 14 and bla CTX-M-27 in dog isolates [7], and bla CTX-M-3 , bla CTX-M-14 , bla CTX-M- 15 , and bla CTX-M-55 , in pig isolates [8]. Thus, domestic animals could shed CTX-M-producing bacteria in their feces. As some domestic animals are reared on the campus, wild weasels may be exposed to AMR bacteria from these animals.
Several genotyping methods, such as pulsed-field gel electrophoresis (PFGE) analysis [7,8], have been used to determine ESBL bacterial transmission based on the genetic relatedness among bacteria [2,4]. However, transfer of ESBL plasmids among bacteria may lead to diversity in the genotypes of ESBL-producing bacteria. Classification of the incompatibility (Inc) type of plasmids is helpful to the epidemiology of ESBL plasmids among several bacterial genotypes [9], and the analysis of plasmid genomes can reveal the history of plasmid transfer [10,11].
This study aimed to clarify the linkage between CTX-M-14 plasmid-carrying E. coli in weasels and other animals present in the facilities located on campus at the Gifu University. Thus, we examined the characteristics and relatedness of the CTX-M-14-producing bacteria and their plasmids.

The Infection Rate and Antimicrobial Susceptibility Profiles of ESBL-Producing E. coli in Animal Facilities and Wild Siberian Weasels
We isolated ESBL-producing E. coli isolates carrying bla CTX-M-14 from weasels (two of nine samples: 22.2%) that were susceptible to non-β-lactam antibiotics. We also isolated ESBL-producing E. coli isolates carrying bla CTX-M-14 from domestic animals (14 of 69 isolates: 20.3%); these isolates were isolated from dog facility 1 (DF1, 8 of 11 samples: 72.7%), dog facility 2 (DF2, four of ten samples: 40%), and cattle facility (CF, two of seven samples: 28.6%, one positive sample from a Holstein cow and another from a beef cow) ( Table 1). The CTX-M-14-producing E. coli isolates from CF, DF1, and DF2 were also susceptible to the non-β-lactams, except for two isolates from DF1 and DF2 that harbored bla TEM-1 as well as bla CTX-M-14 and were resistant to tetracycline-nalidixic acid-ciprofloxacin-chloramphenicol and gentamicin-kanamycin-sulfamethoxazole/trimethoprim, respectively.

PFGE Analysis
We observed seven PFGE profiles in the analyzed CTX-M-14-producing E. coli isolates and classified them into six clusters (clusters A, B, C, D, E, and F), based on 95% pattern similarity from a dendrogram generated from an unweighted pair group method with arithmetic mean (UPGMA) analysis, due to a minor difference of an isolate in cluster C ( Figure 2). Six PFGE profiles from CTX-M-14-producing E. coli isolates were from the fourteen positive fecal samples of domestic animals and one PFGE profile was from the two positive fecal samples from weasels.
Two PFGE profiles were found in DF1, three in DF2, and one in CF. However, the clonal relationship of CTX-M-14-producing E. coli isolates among the various animal facilities or origins was not analyzed in this study.

PFGE Analysis
We observed seven PFGE profiles in the analyzed CTX-M-14-producing E. coli isolates and classified them into six clusters (clusters A, B, C, D, E, and F), based on 95% pattern similarity from a dendrogram generated from an unweighted pair group method with arithmetic mean (UPGMA) analysis, due to a minor difference of an isolate in cluster C ( Figure 2). Six PFGE profiles from CTX-M-14-producing E. coli isolates were from the fourteen positive fecal samples of domestic animals and one PFGE profile was from the two positive fecal samples from weasels.
Two PFGE profiles were found in DF1, three in DF2, and one in CF. However, the clonal relationship of CTX-M-14-producing E. coli isolates among the various animal facilities or origins was not analyzed in this study.

PFGE Analysis
We observed seven PFGE profiles in the analyzed CTX-M-14-producing E. coli isolates and classified them into six clusters (clusters A, B, C, D, E, and F), based on 95% pattern similarity from a dendrogram generated from an unweighted pair group method with arithmetic mean (UPGMA) analysis, due to a minor difference of an isolate in cluster C ( Figure 2). Six PFGE profiles from CTX-M-14-producing E. coli isolates were from the fourteen positive fecal samples of domestic animals and one PFGE profile was from the two positive fecal samples from weasels. Two PFGE profiles were found in DF1, three in DF2, and one in CF. However, the clonal relationship of CTX-M-14-producing E. coli isolates among the various animal facilities or origins was not analyzed in this study.

CTX-M-14 Plasmid Analysis
A total of six plasmids were selected from the six major PFGE profiles of CTX-M-14-producing E. coli isolates from the six clusters. After plasmid-conjugation testing, all transconjugants carrying bla CTX-M-14 with an IncI1 type plasmid and were susceptible to non-β-lactam antibiotics were selected for plasmid DNA extraction. The six plasmids were p130MS from a weasel (cluster A, sample ID: MS7), p105CF from CF (cluster D, sample ID: BC6), p74DF1 and p80DF1 from DF1 (cluster B, sample ID: D11; and cluster F, sample ID: D12, respectively), and p116DF2 and p123DF2 from DF2 (cluster C, sample ID: D23; and cluster E, sample ID: D27, respectively), as shown in Figures 2 and 3. . Physical mapping composition of the sequences of six CTX-M-14 plasmids by BLAST Atlas analysis using p130MS, p105CF, p74DF1, p80DF1, and p116DF2 as query sequences against the longest CTX-M-14 plasmid (p123DF2) as reference sequence. Nucleotide sequences were highly similar among the six plasmids, though p123DF1 had an insertion site. This figure was constructed using the GView Server after performing BLAST Atlas analysis.

Genome Comparisons of the CTX-M-14 Plasmids
After annotating all genes on the sequenced plasmids, the Nucleotide Basic Local Alignment Search Tool (BLASTn) Atlas analysis showed a close similarity in genes among all the plasmids, except for p123DF2, which had a 6940-bp insertion sequence ( Figure 4A). Read mapping analysis, using the CLC Genomics Workbench, showed close nucleotide sequence similarity with plasmids, with differences in two regions located (i) around bla CTX-M-14 (7334 bp) and (ii) within the shufflon system plasmid conjugative transfer pilus tip adhesin gene, PilV (PilV shufflon; 4439 bp; Figure 4).  The differences in nucleotide sequences around bla CTX-M-14 were as follows: (i) p123DF2: 59 bp deleted from IS1380 family transposase (IS1380); 113 bp deleted from the gene encoding phosphotransferase system, histidine-containing phosphocarrier protein; and 6940 bp of six genes inserted within a gene encoding pyruvate formate-lyase protein; (ii) p116DF2: 113 bp deleted from the histidine-containing phosphocarrier protein gene; and (iii) p74DF1: 61 bp deleted from the pyruvate formate-lyase gene. The differences in nucleotide sequences within the PilV shufflon across all plasmids were observed at the 3 -end of the PilV shufflon genes PilVA (p74DF1, p80DF1, and p105CF) and PilVA' (p130MS, p116DF2, p123DF2), and the nucleotide sequence encoded between the PilV shufflon and site-specific recombinase genes.

Relatedness between Six CTX-M-14 Plasmids and Their Interrelatedness with Highly Similar Plasmids
Plasmids p74DF1, p80DF1, p105CF, p130MS, and p116DF2 were the most similar to the CTX-M-1/IncI1 plasmid (GenBank accession number KF787110), with 93% query coverage and 99.63% identity, whereas p123DF2 was also similar to an IncI1 plasmid, but its similar plasmid did not contain any resistance genes (GenBank accession number CP001118) and had 84% query coverage and 98.65% identity. UPGMA analysis showed two clusters of eight plasmids, including a cluster of six CTX-M-14 plasmids used in this study and a cluster of two highly similar plasmids (KF787110 and CP001118; Figure 5A). The UPGMA phylogenetic network of the CTX-M-14 plasmids used in this study showed that a plasmid of DF1 (p80DF1) was closely related to a plasmid from CF (p105CF). Furthermore, the plasmid from CF (p105CF) shared close relatedness with a group of two plasmids from weasel and DF2 (p130MS-p123DF2) and a group of two plasmids from DF1 and DF2 (p74DF1-p116DF2) ( Figure 5B).

Discussion
This study showed clonal spread of CTX-M-14-producing E. coli within each facility on the campus of Gifu University and the linkage of CTX-M-14 plasmids among E. coli isolates from wild weasels and domestic animals on the campus. Siberian weasels are an invasive mammalian predator species in Japan and are distributed in the Gifu Prefecture and adjacent prefectures [12]. The Gifu University campus is located in a rural area of the Gifu Prefecture, which is surrounded by paddy fields, cultivated fields, and rivers. Additionally, the campus has several companion and livestock animal facilities. In this study, CTX-M-14-producing E. coli were isolated in a limited geographical area and carried by free-moving animals (weasels) (Figure 1). A total of three CTX-M-14-producing E. coli-positive facilities (DF1, DF2, and CF) were located within 45 m of each other. However, the PFGE profiles of the isolates from weasels and domestic animals from this study were distinct, suggesting that clonal spread of CTX-M-14-producing bacteria was observed in each facility/origin. The possible lack or failure of biosecurity measures between these facilities in relation with human and animal movements could be the factors responsible for AMR bacterial transfer among these facilities. In addition, wastewater from all facilities was collected by an open drainage system and drained into a main pipeline. Therefore, CTX-M-14 ESBL-encoding plasmids may be transmitted between the bacteria in the limited area of domestic animals after the introduction of ESBL-producing bacteria in each facility.
Some reports have demonstrated closely related plasmids in bacteria with different genotypes [11,13,14]. In this study, CTX-M-14-producing isolates circulating in each animal facility might have harbored a closely related plasmid. The CTX-M-14 plasmids shared identical Inc type, the nearest plasmid sequence type profile with a high percentage of average nucleotide identity, and nucleotide alignment; however, an insertion site in one of the plasmids introduced some differences in plasmid size and other features (Figure 3). The differences in the two genes next to bla CTX-M-14 and the six-gene insertion of p123DF2 near bla CTX-M-14 ( Figure 4A) were associated with the bacterial phosphotransferase system, which is not directly related to plasmid transfer systems [15]. In contrast, differences in the IS1380 gene are related to antimicrobial resistance gene transfer between plasmids. Additionally, the diversity of shufflon segment composition observed in this study ( Figure  4B) could reflect a random recombination between the inverted recombination sites in the shufflon region [16] and/or shufflon rearrangement by the site-specific recombination of genes located next to the shufflon gene [17]. However, no genes encoding proteinshufflon segments were observed in the plasmid genomes, except for shufflon segment D' in p116DF2 ( Figure 4B). Partial IS1380 and formate C-acetyltransferase genes, along with the disappearance of shufflon segments, might have been caused by the nucleotidedeletion site. Evolution of plasmids is necessary to ensure their maintenance in a new bacterial host following transfer from another bacterial cell [18]. In this study, UPGMA analyses revealed close relatedness of the plasmids ( Figure 5), suggesting circulation of the CTX-M-14 plasmid in the studied region. Although the relatedness among plasmids from DF1 and DF2 was observed because of the same UPGMA cluster (Figure 5), the genomes of plasmids from the weasels and DF2 showed close relatedness due to plasmid evolution by insertion/deletion of nucleotide sequences ( Figure 4). Thus, all plasmids had close nucleotide sequence relatedness with differences in the plasmid genome. Although in this study we did not investigate the CTX-M-14 plasmid carriage in other Enterobacteriaceae, there is the possibility that the plasmid can spread to other bacteria in the study location.
AMR bacteria are prevalent in wild animals as well as humans and domestic animals [19]. The wild animal hosts may play a role in maintenance and spread of AMR bacteria and resistance genes. The present results show the linkage between the CTX-M-14 plasmid from domestic animals and wild weasels, although it is not clear whether the transfer of plasmids occurred from domestic animals to wild weasels or vice versa. Antimicrobial drugs are widely used in humans and domestic animals, but not for wild animals. However, wild animals can receive AMR bacteria through the environments close to human activities and animal husbandry and, conversely, transmit the AMR bacteria to humans and domestic animals. In addition, AMR bacteria in wild animals may lead to environmental pollution through their urine and feces. At present, despite low prevalence of AMR bacteria in wild animals in Japan [19], understanding the transmission of AMR bacteria and plasmids among humans, domestic animals, and wild animals is essential for human and animal health. In conclusion, control of the distribution of ESBL-producing bacteria and their ESBL-encoding plasmids is important to prevent the dissemination of ESBL genes to other facilities. In this study, the distribution of CTX-M14-producing bacteria in domestic and wild animals at the Gifu University campus was associated with plasmid transfer.

Samples from Domestic Animals from Animal Facilities
A total of sixty-nine fecal samples were collected between April 2016 and December 2016 from seven animal facilities, including DF1 (n = 11), DF2 (n = 10), (CF, n = 7), and facilities for dairy cows (n = 18), ponies (n = 4), goats (n = 14), and egg-laying hens (n = 5) (Figure 1). Individual fecal samples from each animal were collected from DF1, CF, and the dairy cow and goat facilities. In the remaining facilities, the samples of different animals in the same vicinity (kennels, cages, or stands) were pooled and the final number of samples was as follows: ten from DF2, four from the pony facility, and five from the egg-laying hen facility. DF1, DF2, and CF were open-air facilities located in the same area. DF1 included eleven kennels for healthy mixed breed dogs (one kennel per dog), and all kennels shared the same wastewater drainage system. DF2 included ten kennels for forty healthy beagles (one kennel per four dogs). A total of five kennels of DF2 shared the same wastewater drainage system and walking area. CF was located between DF1 and DF2 and housed two healthy young Japanese black cows in separate enclosures and five healthy Holstein cows in tie stalls. DF1, DF2, and CF shared an equipment storage room (inside CF), a food storage room (inside CF), and a fecal storage barn (located between DF1 and CF). All dogs in DF1 and DF2, and cows in CF, were laboratory animals used by veterinary students and staff for research and education; the students took care of these animals. Dogs in DF1 and DF2 were maintained according to the guidelines on animal research and welfare of Gifu University.
AMR bacteria from the fecal samples were isolated on deoxycholate hydrogen sulfide lactose agar containing 50 µg/mL cephalexin and deoxycholate hydrogen sulfide lactose agar containing 2 µg/mL cefotaxime (CTX; Wako Pure Chemical Industries, Osaka, Japan), and incubated overnight at 37 • C. The isolates were identified at the species level using the API 20E Kit following the manufacturer's instructions.

Ethics Statement
This study was approved by the Ethics Committee for Animal Research and Welfare of Gifu University (approval number 17,109) and by the Ethics Committee for Academic Research of Captured Animals in Gifu Prefecture (approval number 821).

β-Lactamase Gene Identification
β-lactamase genes were identified by performing multiplex polymerase chain reaction (PCR) [23], and the subtypes of the CTX-M-9-group β-lactamases were determined using sequencing analysis with previously reported primer pairs [24]. Both strands of the amplified DNA fragments were sequenced at the Life Science Research Center of Gifu University, and the encoded amino acid sequences were analyzed using the BLAST program (National Center for Biotechnology Information [NCBI], Bethesda, MD, USA).

PFGE Analysis
The genotypes of ESBL-producing E. coli isolates were analyzed according to the standardized PulseNet PFGE protocol for non-O157 E. coli isolates using the CHEF-DR III System (Bio-Rad Laboratories, Hercules, CA, USA); the bacterial genomic DNA was digested with the restriction enzyme XbaI (Takara Bio, Shiga, Japan). XbaI-digested DNA of the Salmonella serotype Braenderup H9812 isolate was used as the universal size standard. The PFGE profiles were analyzed using a UPGMA dendrogram with 1% optimization and 1% band-filtering tolerance with BIONUMERICS software version 7.6.1 (Applied Maths NV, Sint-Martens-Latem, Belgium).

Selected Isolates for Plasmid Sequencing
After PFGE analysis, plasmid characterization was performed. At least one isolate per cluster (based on PFGE analysis) was selected and subjected to plasmid-conjugation using the broth mating method with rifampicin-and nalidixic acid-resistant E. coli DH5α as recipients as described previously [25]. The transconjugants were subjected to PCR analysis using primers for the CTX-M-9 group [24], and antimicrobial susceptibility analysis and PCR-based plasmid replicon typing using previously reported primer sets [9] were performed.
A transconjugant isolate with an Inc type plasmid was selected from each cluster for plasmid DNA purification. Briefly, plugs of the transconjugants were incubated with the S1 Nuclease enzyme (Promega, Madison, WI, USA) and subjected to PFGE in a SeaKem Agarose gel (Lonza, Rockland, ME, USA) [26]. After staining the S1-PFGE gel with SYBR Gold Nucleic Acid Stain (Invitrogen, Carlsbad, CA, USA), the plasmid bands were cut with the aid of a blue-light transilluminator and purified using the MonoFas DNA Purification Kit (GL Sciences, Tokyo, Japan). Purified plasmid DNA was sequenced on the MiSeq System (Illumina, San Diego, CA, USA) according to the manufacturer's instructions using the Nextera XT Library Prep Kit (Illumina), optimized for plasmids. De novo assembly was performed using the A5-miseq pipeline [27], followed by manual annotation with the GPAT, available in the database of pathogen genomic and epidemiology (GenEpid-J) (Pathogen Genomic Center, National Institute of Infectious Diseases, Toyama, Japan). The plasmid Inc and sequence types were investigated using GPAT. To identify plasmid similarity, BLAST was used to calculate the average nucleotide identity and perform nucleotide alignments of the plasmid sequences [28]. The percentage of aligned nucleotides was calculated as [(number of similar nucleotides between query and reference sequences after alignment)/(number of reference sequences)] × 100. The plasmid genomes were compared by visualization with CLC Genomics Workbench software version 10.0.1 (Qiagen, Hilden, Germany). A BLAST Atlas view was generated using the GView server [29]. Genome mapping analysis was performed using Multiple Alignment of Conserved Genomic Sequence with Rearrangements (Mauve) [30] and the GView program version 1.7 [29].

Relatedness among All Plasmids Used in this Study and Interrelatedness between the Highly Similar Plasmids
All plasmid genomes used in this study were queried for highly similar sequences using Standard Nucleotide BLAST of the NCBI database. The plasmid genomes used in this study, and their most highly similar sequences, were aligned and imported into SplitsTree software version 4.11.3 (SplitsTree4) [31] to construct the UPGMA phylogenetic network for analyzing plasmid relatedness. Thereafter, only the plasmid genome sequences of interest were aligned and imported into SplitsTree4 to construct the UPGMA phylogenetic network for analyzing the relatedness among them.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/antibiotics10040432/s1. Table S1: average nucleotide identity (%) and aligned nucleotides (%, number of aligned nucleotides between query and reference sequences/bases in length of reference sequence] of six plasmid sequences.