Comparison of IncK-blaCMY-2 Plasmids in Extended-Spectrum Cephalosporin-Resistant Escherichia coli Isolated from Poultry and Humans in Denmark, Finland, and Germany

Escherichia coli carrying IncK-blaCMY-2 plasmids mediating resistance to extended-spectrum cephalosporins (ESC) has been frequently described in food-producing animals and in humans. This study aimed to characterize IncK-blaCMY-2-positive ESC-resistant E. coli isolates from poultry production systems in Denmark, Finland, and Germany, as well as from Danish human blood infections, and further compare their plasmids. Whole-genome sequencing (Illumina) of all isolates (n = 46) confirmed the presence of the blaCMY-2 gene. Minimum inhibitory concentration (MIC) testing revealed a resistant phenotype to cefotaxime as well as resistance to ≥3 antibiotic classes. Conjugative transfer of the blaCMY-2 gene confirmed the resistance being on mobile plasmids. Pangenome analysis showed only one-third of the genes being in the core with the remainder being in the large accessory gene pool. Single nucleotide polymorphism (SNP) analysis on sequence type (ST) 429 and 1286 isolates showed between 0–60 and 13–90 SNP differences, respectively, indicating vertical transmission of closely related clones in the poultry production, including among Danish, Finnish, and German ST429 isolates. A comparison of 22 ST429 isolates from this study with 80 ST429 isolates in Enterobase revealed the widespread geographical occurrence of related isolates associated with poultry production. Long-read sequencing of a representative subset of isolates (n = 28) allowed further characterization and comparison of the IncK-blaCMY-2 plasmids with publicly available plasmid sequences. This analysis revealed the presence of highly similar plasmids in ESC-resistant E. coli from Denmark, Finland, and Germany pointing to the existence of common sources. Moreover, the analysis presented evidence of global plasmid transmission and evolution. Lastly, our results indicate that IncK-blaCMY-2 plasmids and their carriers had been circulating in the Danish production chain with an associated risk of spreading to humans, as exemplified by the similarity of the clinical ST429 isolate to poultry isolates. Its persistence may be driven by co-selection since most IncK-blaCMY-2 plasmids harbor resistance factors to drugs used in veterinary medicine.


Introduction
Extended-spectrum cephalosporins (ESC) are classified as critically essential antimicrobials by the World Health Organization.Therefore, their use is restricted to the treatment of infections in humans and animals caused by multidrug-resistant (MDR) Gram-negative bacteria, particularly Enterobacteriaceae such as MDR Escherichia coli (E.coli) [1].In Europe, the Antimicrobial Advice Ad Hoc Expert Group (AMEG) of the European Medicines Agency (EMA) has classified third-and fourth-generation cephalosporins as belonging to "Category B-Restrict" antibiotics, meaning their use in animals should be restricted and preferably be based on antimicrobial susceptibility testing to mitigate the risk to public health [2].Resistance to ESCs has dramatically increased worldwide in this family of bacteria and is encoded predominantly by the extended-spectrum β-lactamase (ESBLs) and AmpC β-lactamase (AmpCs) genes [3,4].The plasmid-mediated AmpC-like gene, bla CMY-2 , which reduces susceptibility to aztreonam, cephamycin, and third-generation cephalosporins, is one of the most prevalent among E. coli strains [5].In general, the gene bla CMY-2 is located on plasmids belonging to several incompatibility groups, including IncI1, IncA/C, IncF, and IncK [6][7][8][9][10].
The incidence of bla CMY-2 -related ESC resistance in E. coli belonging to various multilocus sequence types (STs) is common in livestock across several European countries, while being rarely reported in E. coli from humans in Europe [11][12][13].Previous studies observed that IncI1 and IncK plasmids carrying bla CMY-2 occurred in ST38, ST131, and ST117 ESC-resistant E. coli isolates from human, livestock, and meat products in Germany, as well as in the same E. coli STs from healthy urban dogs in France [14,15] Also, IncK and IncI plasmids harboring bla CMY-2 were identified in Finnish poultry farms [16], while IncA/C plasmids containing bla CMY-2 were detected in particular E. coli lineages in Swedish broiler flocks [17].Surveillance of antimicrobial resistance in Denmark in 2015-2016 documented that E. coli isolates carrying bla CMY-2 in Danish broiler meat came from widely distributed STs, i.e., ST38, ST154, ST2309, while imported broiler meat isolates belonged to ST23, ST117, ST131, ST2040 [18].A recent study from Denmark revealed a close phylogenetic relatedness among E. coli ST131 IncI1-bla CMY-2 plasmids carrying isolates from broilers and a patient with a bloodstream infection which highlights the risk of the potential zoonotic spread of these antimicrobial-resistant bacteria [19].A novel ST429 ESC-resistant E. coli harboring bla CMY-2 was detected in the Danish surveillance of both domestic broilers and their meat, in imported broiler meat [18] as well as in our previous study, which revealed that bla CMY-2 on IncI (ST2040) or IncK (ST429) plasmids dominated in a survey of ESC-resistant E. coli from Danish poultry farms and slaughterhouses over the period 2015-2018 [20].However, knowledge is currently lacking regarding the IncK-bla CMY-2 plasmids' relatedness among different ESC-resistant E. coli STs, including ST429, in the poultry production chain in Denmark and other countries, and from human clinical cases.
Therefore, the objective of this study was to investigate the genetic diversity of IncKbla CMY-2 -positive ESC-resistant E. coli isolates (n = 46) and further to identify the similarity of IncK-bla CMY-2 plasmids (n = 28) from the poultry production chains in Denmark, Finland, and Germany as well as from Danish patients.Also, we determined the spread of ST429 ESC-resistant E. coli carrying bla CMY-2 globally and compared IncK-bla CMY-2 plasmids using publicly available databases.

Phenotypic Antimicrobial-Resistance Profiles
Results of MIC distributions are summarized in Table 1 for all ESC-resistant E. coli IncK-bla CMY-2 isolates.Specific results for each of the isolates are provided in Supplementary Table S1.A total of 15 resistance profiles were observed.Based on the epidemiological cut-off values (ECOFFs), all 46 isolates were classified as non-wild type (i.e., resistant to antibiotics) to ampicillin, cefotaxime, and ceftazidime.Moreover, 93.5%, 84.8%, and 71.7% of the ESC-resistant E. coli isolates were classified as resistant to sulfamethoxazole, tetracycline, and gentamicin.In addition, the proportion of isolates classified as non-wild type to trimethoprim, nalidixic acid, and ciprofloxacin ranged from 19.6% to 8.6%.For other antimicrobials, MICs above ECOFFs were less common to azithromycin (2.2%) and chloramphenicol (2.2%), and all isolates were wild type with respect to amikacin, colistin, and tigecycline.Overall, all ESC-resistant E. coli isolates showed multidrug resistance, as they exhibited resistance towards three or more antibiotics belonging to different classes [21].The % of isolates is for antimicrobial resistance within the range under study.When the growth of isolates was observed at the highest concentration of the antibiotic tested, the MIC value was recorded as the following concentration (e.g., for ampicillin, 97.8% of the isolates exhibited visible growth at a concentration of 32 mg/L, revealing an MIC of >32 mg/L and an annotation of the final MIC as 64 mg/L).

Results from Conjugation Assays
The conjugation experiments showed that all 46 ESC-resistant E. coli isolates could transfer cefotaxime resistance to the sensitive E. coli MG1655 with PCR results confirming the presence of the blaCMY-2 gene in the transconjugants.These results indicate that the blaCMY-2 gene in all ESC-resistant E. coli isolates regardless of their ST-types was located on a conjugative plasmid (Table 2).

Results from Conjugation Assays
The conjugation experiments showed that all 46 ESC-resistant E. coli isolates could transfer cefotaxime resistance to the sensitive E. coli MG1655 with PCR results confirming the presence of the bla CMY-2 gene in the transconjugants.These results indicate that the bla CMY-2 gene in all ESC-resistant E. coli isolates regardless of their ST-types was located on a conjugative plasmid (Table 2).[22].e -ESC-resistant E. coli isolates from the German chicken meat, collected in a previously published study [14].

Pangenome Analysis
Pangenome analysis accounts for the genomic flexibility among the 46 IncK-bla CMY-2positive ESC-resistant E. coli isolates by considering the core and pangenome structure of the isolates, as shown in Figure 2. The pangenome in Figure 2a consisted of 11,804 genes in total, which could be divided into core genes, soft-core genes, shell genes, and cloud genes.The core genome (3094 genes, 27.91%) made up less than one-third of the collective pangenome genes for the ESC-resistant E. coli isolates; in contrast, results from the Roary pipeline generated 7990 accessory genes in total showing large variability among the isolates.Of these genes, soft-core genes found in 95-99% of the isolates comprised 351 genes (3.17%).Shell genes (2339 genes, 21.10%) were detected in between 15 and 95% of the isolates.Cloud gene families (5300 genes, 47.82%) were represented in less than 15% of the isolates, highlighting the high genomic variability.The pangenome clustering tree in Figure 2b shows the relatedness among ST429 isolates from different sources and countries of origin vs. the other STs, e.g., 57, 162, and 1286 (see Table 2 for information on ST), found in a separate cluster.Figure 2c visualizes the presence and absence of annotated genes in each of the isolates, clearly showing the shared core genes and the range of shell and clouds genes.Further details are found in Supplementary Table S2.

Pangenome Analysis
Pangenome analysis accounts for the genomic flexibility among the 46 IncK-blaCMY-2positive ESC-resistant E. coli isolates by considering the core and pangenome structure of the isolates, as shown in Figure 2. The pangenome in Figure 2a consisted of 11,804 genes in total, which could be divided into core genes, soft-core genes, shell genes, and cloud genes.The core genome (3094 genes, 27.91%) made up less than one-third of the collective pangenome genes for the ESC-resistant E. coli isolates; in contrast, results from the Roary pipeline generated 7990 accessory genes in total showing large variability among the isolates.Of these genes, soft-core genes found in 95-99% of the isolates comprised 351 genes (3.17%).Shell genes (2339 genes, 21.10%) were detected in between 15 and 95% of the isolates.Cloud gene families (5300 genes, 47.82%) were represented in less than 15% of the isolates, highlighting the high genomic variability.The pangenome clustering tree in Figure 2b shows the relatedness among ST429 isolates from different sources and countries of origin vs. the other STs, e.g., 57, 162, and 1286 (see Table 2 for information on ST), found in a separate cluster.Figure 2c visualizes the presence and absence of annotated genes in each of the isolates, clearly showing the shared core genes and the range of shell and clouds genes.Further details are found in Supplementary Table S2.

Phylogenetic Tree Analysis
The SNP-based trees revealed close relationships among isolates belonging to ST429 and ST1286, respectively.Isolates were obtained from different sources, i.e., poultry, poultry meat, and humans, over a long period from 2012 to 2018.As shown in Figure 3a, ST429 isolates obtained in a Danish slaughterhouse (SA17021, SA17022) and one human clinical case (HBI01) were highly similar with only 4-7 SNPs, indicating possible transfer from poultry to humans.Notably, two ST429 isolates (SA17023, SA17024), collected from the same slaughterhouse in Denmark, belonged to the same subcluster as the three German isolates from 2012 and all seven Finnish isolates also obtained in 2017, which suggests the introduction of the ST subcluster from a common source and clonal transmission in the poultry production chain.Furthermore, German isolates from 2012 clustering with Danish poultry and one human clinical isolate from 2015-2017, with 51 SNPs between the German RL97 and the Danish SA17023, again point out the longevity of this ST429 in chicken production and possible common sources transcending country boundaries.All six ST1286 isolates in Figure 3b came from one slaughterhouse in Denmark between 2017 and 2018 and differed by SNPs ranging from 13 (SB17103 vs. SB18102) to 90 (SB18001 vs. SB18004), suggesting a temporal colonization or repeated introduction of ST1286 in the Danish slaughterhouse.The resulting SNP matrices of ST429 and ST1286 are shown in Supplementary Table S3.
found in <15% of the E. coli in the study.(b) Maximum likelihood phylogenetic tree inferred from the alignment of the 3094 core genes of 46 ESC-resistant E. coli by FastTree.(c) Annotation of gene presence (blue) and absence (white) matrix across the pangenome of the E. coli.The top scale shows the complete genome size (kbp).Each row shows the gene content of an E. coli isolate.Each column shows the comparative gene clusters.The data were visualized using Phandango.

Phylogenetic Tree Analysis
The SNP-based trees revealed close relationships among isolates belonging to ST429 and ST1286, respectively.Isolates were obtained from different sources, i.e., poultry, poultry meat, and humans, over a long period from 2012 to 2018.As shown in Figure 3a, ST429 isolates obtained in a Danish slaughterhouse (SA17021, SA17022) and one human clinical case (HBI01) were highly similar with only 4-7 SNPs, indicating possible transfer from poultry to humans.Notably, two ST429 isolates (SA17023, SA17024), collected from the same slaughterhouse in Denmark, belonged to the same subcluster as the three German isolates from 2012 and all seven Finnish isolates also obtained in 2017, which suggests the introduction of the ST subcluster from a common source and clonal transmission in the poultry production chain.Furthermore, German isolates from 2012 clustering with Danish poultry and one human clinical isolate from 2015-2017, with 51 SNPs between the German RL97 and the Danish SA17023, again point out the longevity of this ST429 in chicken production and possible common sources transcending country boundaries.All six ST1286 isolates in Figure 3b came from one slaughterhouse in Denmark between 2017 and 2018 and differed by SNPs ranging from 13 (SB17103 vs. SB18102) to 90 (SB18001 vs. SB18004), suggesting a temporal colonization or repeated introduction of ST1286 in the Danish slaughterhouse.The resulting SNP matrices of ST429 and ST1286 are shown in Supplementary Table S3.S4).The Danish ST429 bla CMY-2 -positive isolates were mainly found in two closely related clusters with isolates collected from other countries, with allelic differences ≤ 15.To note, the Danish isolates were observed to be highly similar to isolates (at the center of the tree) originating from UK poultry between 2013 and 2016, suggesting the possibility that ST429 bla CMY-2 -positive Danish isolates were initially imported from Britain.Interestingly, SA17023 (Supplementary Table S4) grouped into the same subcluster with the two Sweden poultry isolates ESC_RA0430AA (accession no.SRR11473342) and ESC_RA0434AA (accession no.SRR11473346) in 2016 with only three SNP differences, and PS16004 showed five allelic differences with one isolate ESC_UA5215AA (accession no.ERR5443272) from Spain poultry in 2016.Meanwhile, these two subclusters with SA17023 and PS16004 also included UK poultry isolates from 2015, indicating that the poultry parent birds of Denmark, Sweden, and Spain possibly were imported from Britain.
and Germany and 1 human isolate from Denmark are shown in black bold.Isolates from the same date and source are presented in the same color.The observed number of SNPs among these isolates is indicated as well.
The minimum spanning tree based on the 102 ST429 ESC-resistant E. coli isolates from this study (Danish isolates 2015-2017, n = 22) and Enterobase (n = 80, 2010-2021) is shown in Figure 4. Overall, our results demonstrated the worldwide spread of ST429 ESCresistant E. coli over the past decade, suggesting a role for global trade in the dissemination of these isolates.Of the 102 isolates depicted in Figure 4, 84 were related to poultry production (Supplementary Table S4).The Danish ST429 blaCMY-2-positive isolates were mainly found in two closely related clusters with isolates collected from other countries, with allelic differences ≤ 15.To note, the Danish isolates were observed to be highly similar to isolates (at the center of the tree) originating from UK poultry between 2013 and 2016, suggesting the possibility that ST429 blaCMY-2-positive Danish isolates were initially imported from Britain.Interestingly, SA17023 (Supplementary Table S4) grouped into the same subcluster with the two Sweden poultry isolates ESC_RA0430AA (accession no.SRR11473342) and ESC_RA0434AA (accession no.SRR11473346) in 2016 with only three SNP differences, and PS16004 showed five allelic differences with one isolate ESC_UA5215AA (accession no.ERR5443272) from Spain poultry in 2016.Meanwhile, these two subclusters with SA17023 and PS16004 also included UK poultry isolates from 2015, indicating that the poultry parent birds of Denmark, Sweden, and Spain possibly were imported from Britain.Isolates are colored according to country and year of isolation indicated in circles (each circle may contain more than one isolate in cases of no allelic differences).The number of isolates per country is indicated in brackets in the color legend.Numbers of allelic differences between isolates are shown in the connecting lines.More detailed information on isolates and the minimum spanning trees can be found in Supplementary Table S4 and Figure S3a-d.

Plasmid Characterization
The short-read sequencing and subsequent hybrid assembly (long-and short-reads combined) of plasmids from a subset of 28 representative E. coli bla CMY-2 isolates belonging to nine STs revealed that all isolates harbored IncK plasmids carrying bla CMY-2 , with predicted sizes ranging from 85,938 to 133,315 bp (Table 2).Also, all hybrid assembled IncK-bla CMY-2 plasmids were found to be in one circular contig.Resfinder confirmed that six IncK plasmids harbored only the bla CMY-2 resistance gene, i.e., plasmids pSB17032 and pSB17042 from poultry samples, and pHBI02, pHBI03, pHBI04, and pHBI05 from clinical isolates.Plasmids present in all ST429 and ST162 isolates in this study contained the additional resistant genes sul1, aac (3)-Via, aadA1, and tet(A), except for plasmid pPS15001 (tetA was absent).We also found that six IncK plasmids from ST1286 isolates carried bla TEM-1 and bla CMY-2 , which co-occurred with the dfrA1, sul2, and tet(A) resistant genes.VirulenceFinder detected traT (n = 28, 100%) and cib (n = 22, 78.6%) in all or most plasmids.

Comparison of IncK-bla CMY-2 Plasmids from This Study and Public Databases
A comparison of the IncK-bla CMY-2 plasmids under study showed that all have the conjugative transfer genes encoding pilus (pilI, pilP, pilR, pilS, pilM, pilN, and pilV), genes from the transference operon tra (traC-F, traH, traJ, traM-T, traV-W, and traX-Y), and genes encoding for a DNA primase (T7 DNA primase) and an endonuclease relaxase (nikB) (Figure 5 and Supplementary Figure S1).They also shared genes involved in plasmid partition/stability.Interestingly, four plasmids of poultry origin belonging to ST429 (n = 3) and ST162 (n = 1), namely, pSB17040, pRS184-R, pRL97, pSA17101, and the plasmid pHBI01 from a human blood ST429 isolate, showed 99.9% identity: they displayed common virulence factors and resistance genes (Table 2) although their origin (production systems, slaughterhouses, or clinical, as well as years and countries of isolation) differed.The plasmid pPS15001, found in an E. coli ST429, lacked a 10 kb region with tet(A) and mercury-resistance genes compared to the remaining plasmids linked to ST429 (Table 2).Importantly, ST1286 E. coli isolates emerging in 2018 contained the largest IncK-bla CMY-2 plasmid (132-133 kb) that harbored some unique genes encoding a transposase, the toxinantitoxin system (relE/parE), and the SOS inhibition genes psiAB not found in the other IncK-bla CMY-2 plasmids.However, the mercury-resistance operon (mer) was not detected in IncK-bla CMY-2 plasmids from E. coli ST1286 (Supplementary Table S5).IncK-bla CMY-2 plasmids from ST429 E. coli included in this study exhibited a high similarity with publicly available IncK plasmids present in E. coli ST429 from Norway and the USA (Supplemental Material Figure S2), while a Norwegian IncK-bla CMY-2 plasmid from an E. coli ST162 isolate lacked a 45 kb fragment compared to the Danish pSA17101 from E. coli ST162.In addition, plasmids pSB17032 and pSB17042, from E. coli ST57 and ST350 isolated from slaughterhouses, were almost identical to the pHBI003 plasmid from the Danish human blood E. coli ST131 isolate, as shown in Figure 5.
The phylogenetic tree with 53 IncK-bla CMY-2 complete sequences from the GenBank database (Supplementary Table S6) and 28 IncK-bla CMY-2 plasmids from the current study is shown in Figure 6.Among these plasmids, 75 plasmids, showing a worldwide distribution, harbored only one β-lactamase gene, bla CMY-2 , while six plasmids, restricted to Denmark, also contained the bla TEM-1 resistance gene.The plasmid sizes ranged from 70 kb to 133 kb.The conjugation-associated traT gene was found in 79 (98%) plasmids, while 40 plasmids showed, in addition, the virulence factor colicin Ib (cib gene), which is a polypeptide toxin acting against E. coli and closely related bacteria.Phylogenetic analysis showed that three plasmids, pPS15001 from Denmark, p30P2 from the USA (accession no.LC557961.1),and p22C121-2 from Japan (accession no.LC501554.1),containing the same additional antimicrobial-resistance (aac (3)-Via, aadA1, sul1) and virulence genes (cib, traT) clustered with plasmids that, besides the above-mentioned genes, also carried tet(A), indicating a putative common ancestor and subsequent introduction of the tet(A) gene.The IncK-bla CMY-2 plasmid pN16S065 (accession no.CP082750.1)recovered from a Salmonella enterica isolate in the USA clustered with plasmids from E. coli, suggesting the spread of these plasmids to other Enterobacteriaceae.Moreover, IncK-bla CMY-2 plasmids isolated from human clinical cases were interspersed with plasmids from animals, food, and the environment, suggesting that IncK-bla CMY-2 plasmids and their hosts circulate in the entire ecosystem.

Discussion
This study characterized the genetic diversity of 46 IncK-blaCMY-2-positive ESCresistant E. coli isolates previously collected from the Danish poultry production system and human clinical samples during 2015-2021, Finnish broilers in 2017, and German chicken meat in 2012.Using public databases and data from the present study, the epidemiology of ST429 E. coli isolates and IncK-blaCMY-2 plasmids in the context of poultry production and possible links to clinical cases was evaluated.The results indicated that genetically diverse ESC-resistant E. coli STs with IncK-blaCMY-2 plasmids have been circulating in the poultry production chain in Denmark and other countries for over a decade.Moreover, this study strongly suggests that transfers to humans are occurring

Discussion
This study characterized the genetic diversity of 46 IncK-bla CMY-2 -positive ESC-resistant E. coli isolates previously collected from the Danish poultry production system and human clinical samples during 2015-2021, Finnish broilers in 2017, and German chicken meat in 2012.Using public databases and data from the present study, the epidemiology of ST429 E. coli isolates and IncK-bla CMY-2 plasmids in the context of poultry production and possible links to clinical cases was evaluated.The results indicated that genetically diverse ESC-resistant E. coli STs with IncK-bla CMY-2 plasmids have been circulating in the poultry production chain in Denmark and other countries for over a decade.Moreover, this study strongly suggests that transfers to humans are occurring with links to blood infections.Previous studies have also documented the spread of IncK-bla CMY-2 plasmids in ESC-resistant E. coli isolates obtained from 2006 to 2012 across different host species, including humans in Denmark [23].
All the bla CMY-2 -positive E. coli isolates carry multiple resistance genes, with the in silico results agreeing with the phenotypic analysis (Figure 2, Table 1).The gentamicin-resistance genes aac (3)-Vla and aac (3)-IId were frequently detected in the IncK-bla CMY-2 -positive ESC-resistant E. coli isolates, genes also observed in E. coli from chicken and human sources in Canada [29].
One human isolate (HBI02) harbored both the azithromycin-resistance gene mph(A) and the chloramphenicol-resistance gene catA1.Co-harborage of the bla CMY-2 , mph(A), and catA1 genes was previously reported in E. coli isolates from calves in the USA [30].A recent study from Korea detected the catA1 gene in bla CMY-2 -producing pathogenic E. coli in pigs, but there were no chloramphenicol-resistant genes in the strains isolated from humans [31].In addition, four of the IncK-bla CMY-2 -positive bla CMY-2 -positive E. coli isolates showed chromosomal point mutations yielding fluoroquinolone resistance (Figure 1).These isolates had at least two mutations in the gyrA gene combined with high MIC values of ciprofloxacin (0.25-0.5 mg/L) and simultaneously nalidixic acid MIC values ≥ 128 mg/L.Single mutations in the gyrA gene altering leucine at position 83 (S83L), as well as double mutations in the same gene altering positions at aspartic acid (D87N) and serine (S83L), are known to turn E. coli strains resistant to fluoroquinolones [32].
All 46 IncK-bla CMY-2 -positive ESC-resistant E. coli could transfer the bla CMY-2 gene by conjugation to a recipient, pointing to the potential for horizontal transmission of the plasmid-borne bla CMY-2 gene.A previous study in Denmark identified that exogenous E. coli of human or animal origin could readily transfer bla CMY-2 -encoding plasmids to the human fecal microbiota [33].Furthermore, an IncK-bla CMY-2 plasmid was transferable between E. coli and S. Heidelberg isolates, but the transfer was unsuccessful between S. Heidelberg isolates, as described by [34].Also, a previous study demonstrated the presence of IncKbla CMY-2 plasmids in Salmonella enterica in the USA [35], plasmids which interestingly were similar to the ones detected in the present study (Figure 6).
The pangenome analysis revealed that the IncK-bla CMY-2 E. coli isolates possess a large source gene pool and the capacity to acquire novel genetic elements.The pool of conserved core genes is three times smaller than the pools of accessory and cloud genes, which suggests a flexible genome [36].Other studies have also described the core genome of E. coli as being comparatively small [37][38][39].However, it is worth emphasizing that the core genome is relative, as the concatenated core would become smaller if more genomes were added to the comparison [40].On the other hand, it is notable that the smaller size of core genomes will result in more expansive accessory genomes and isolate-specific cloud genes [41].Of course, these non-essential accessory and cloud genes are prone to rapid evolution, and their knockout does not impact the isolate phenotype [42].In addition, the tree provides deeper insight into the concatenated core gene alignment, which indicates that various E. coli STs have been of great significance in the evolution of IncK-bla CMY-2 E. coli isolates, as previously described [43].
IncK-bla CMY-2 -positive ST429 E. coli isolates with 0-54 SNP differences were originally isolated from various stages of poultry production and from humans, implying that clonal transmission happens between different hosts.ST429 isolates from Finland and Germany included in this study were previously demonstrated to show few SNP differences in each case/country [14,22].In addition, ST1286 isolates with 13-90 SNP differences were originally collected from slaughterhouses [20], also suggesting clonal transmission.A recent study from China reported no SNP differences among ST1286 E. coli isolates from laying hens [44].While ST1286 IncK-bla CMY-2 -positive E. coli isolates were found in poultry, this ST was not among our IncK-bla CMY-2 -positive ESC-resistant E. coli isolates from humans, indicating lower virulence potential, in agreement with previous studies [45,46].
The cgMLST analysis of ST429 E. coli isolates from Denmark [20] and Enterobase (n = 81) showed that the isolates differed in a limited number of alleles, highlighting the existence of a conserved pool of ST429 carrying bla CMY-2 and supporting the transmission of bla CMY-2 along the poultry production chain and across sectors.Moreover, the cgMLST analysis of ST429 E. coli isolates carrying other bla-family genes such as bla CTX-M also identified the clonal relationship between isolates from different countries [47].In addition, a recent pan-European study reported ESC-resistant E. coli isolates carrying bla CMY-2 on different Inc plasmid types (e.g., IncI1, IncK2, IncA/C) in diverse STs, reflecting the dissemination of cephalosporin-resistance genes via successful plasmid lineages [48].
IncK plasmids of different sizes and genetic contents were demonstrated by short-read and long-read sequencing to contain a conserved carrier/location of the bla CMY-2 gene and to occur in genetically diverse STs.IncK plasmids can be divided into two separate lineages, namely, IncK1, which is often associated with bla CTX-M-14 , and IncK2, which predominately carries bla CMY-2 [49,50].Previous studies have described, using hybrid assembly, transferable IncK-bla CMY-2 plasmids in ESC-resistant E. coli isolates in other countries [51,52].Here, we found that one ST429 Danish E. coli harbored a smaller (~109 kb) IncK-bla CMY-2 plasmid, while the remaining eight ST429 isolates from Denmark harbored a longer version (119-120 kb).Thus, our results suggest that at least two variants of IncK-bla CMY-2 plasmids are circulating among ST429 ESC-resistant E. coli isolates in the poultry production chain in Denmark.To note, the most common Danish IncK-bla CMY-2 plasmids were highly similar to plasmids originating from broiler production in other countries, IncK-bla CMY-2 plasmids circulating in Germany (117-120 kb) in 2012, in Finland (~122 kb) in 2017 [14,43], and in Norway (~110 kb) [53], indicating the successful spread of IncK-bla CMY-2 plasmids in broiler production.Also, highly similar IncK-bla CMY-2 plasmids linked to three different E. coli STs from two slaughterhouses were observed, indicating a potential horizontal dissemination among E. coli STs in broiler production, as suggested by others [54].
Among the five IncK-bla CMY-2 -positive E. coli clinical isolates linked to different STs, one clinical isolate of E. coli ST429 harbored a IncK-bla CMY-2 plasmid (pHBI01), which shared high sequence homology with those from E. coli ST429 from poultry (Figure 5), providing evidence of the vertical clonal spread of E. coli harboring IncK-bla CMY-2 plasmids between poultry and humans.While E. coli ST429 isolates represent a common avian pathogenic lineage specific to poultry, it was previously believed to hold little pathogenic potential for humans [55].However, extraction of E. coli ST429 isolates from Enterobase revealed the presence of eight isolates harboring IncK-bla CMY-2 plasmids, which were derived from human clinical cases during 2016-2021 (Supplementary Table S4).Taken together, this may imply a stronger pathogenicity potential for this ST than previously thought.
To note, the IncK-bla CMY-2 plasmids pHBI02 and pHBI03, linked to E. coli ST69 and ST131 isolates from humans, exhibit a high similarity with IncK-bla CMY-2 plasmids related to E. coli ST350 and ST57 from slaughterhouse meat (Figure 5), indicating a horizontal transfer of plasmids from poultry to clinical E. coli lineages in Denmark.E. coli ST131 is responsible for 50% of ESBL blood infections with no recognized animal reservoir [56], although a link to broilers was implied in a 2019 study [19].E. coli ST69 is a globally distributed E. coli responsible for hospital-acquired antimicrobial-resistant human infections [57] and known to be able to colonize the animal intestine [58].E. coli ST1286 (n = 6 strains), according to DANMAP [59], is not a frequent ST associated with bla CMY-2 in broilers or broiler meat, which agrees with our previous study [20], where ST1286 isolates carrying bla CMY-2 were only isolated from poultry meat from one slaughterhouse in 2017-2018.
The detailed plasmid comparison performed in this study revealed a common IncKbla CMY-2 backbone sequence of 85 kb for all 28 plasmids.This plasmid backbone sequence might be common in poultry, as described earlier [50,54].Additionally, most of our IncKbla CMY-2 plasmids contain toxin-antitoxin systems (e.g., pndC/ydfB or relE/ParE) and regions involved in plasmid transfer (such as tra, pil operons) that ensure stable maintenance of the plasmid to the daughter cells after cell division and the spread to other strains, respectively, which increases the potential for transmission to opportunistic pathogens in the poultry and human gut microbiota or in environmental reservoirs [60].Interestingly, a mercury-resistance operon (mer, [61]) was observed in IncK-bla CMY-2 plasmids from E. coli ST429 originally isolated from poultry and clinical samples in Germany, Finland, and Denmark [14,20,43], whereas mer genes were not detected in IncK-bla CMY-2 -positive E. coli from other STs (poultry isolates, ST1286, ST162, ST350, and ST57; and clinical isolates, ST69, ST131, ST95, and ST12).It is possible that contamination of poultry or other animal feed in the farms promoted the carriage of mercury genes in IncK-bla CMY-2 plasmids.Along these lines, mercury was detected in mineral feed used in poultry rearing in Germany in 2013 [62].
In this study, we created a phylogenetic tree of IncK-bla CMY-2 plasmids composed of sequences from this study (n = 28) and publicly available sequences at GenBank, NCBI (n = 53, from seven other countries).To our knowledge, this is the first attempt to compare all IncK-bla CMY-2 plasmids published so far (n = 81).Results highlight that IncK plasmids represent a major vehicle for bla CMY-2 and other antimicrobial-resistance genes worldwide and over time, overall, across the poultry production chain but also in humans.E. coli was the dominant host with only one IncK-bla CMY-2 plasmid carried by Salmonella enterica [63].Recent studies have reported that bla CMY-2 is also harbored by IncI1 plasmids present in Salmonella enterica isolates from chicken meat in Spain and South Korea [64,65].Since 39 (48.1%)IncK-bla CMY-2 plasmids were confirmed to carry genes encoding resistance to sulfonamides (n = 35) and tetracyclines (n = 25) all over the world, the spread of such plasmids might be prompted by the use of these antimicrobial agents worldwide.A total of 79 (97.5%)IncK-bla CMY-2 plasmids harbored the traT gene and 40 (49.4%)contained the cib genes which indicates the horizontal transfer of the IncK-bla CMY-2 plasmid and co-location of bla CMY-2 with virulence factor colicin Ib, respectively [66].A previous study showed that the mutation of the traY gene of IncK plasmids effectively prevents conjugation [67].

Plasmid Conjugation
Transfer of the plasmids was confirmed in a filter conjugation mating assay with some modifications [73].All bla CMY-2 carrying ESC-resistant E. coli isolates were used as donors, and E. coli MG1655 resistant to rifampicin and nalidixic acid served as a recipient strain.Donor and recipient strains were cultured overnight in Luria-Bertani broth (LB, Sigma-Aldrich, St. Louis, MO, USA) at 37 • C and washed twice with phosphate-buffed saline (PBS, Invitrogen, Maryland, MD, USA).After adjusting the OD600 to 0.5, donor and recipient isolates were mixed at a ratio of 1:1, and 100 µL were immediately applied to a 0.2 µm nitrocellulose filter membrane, placed on LB agar, followed by overnight culture at 37 • C for 20 h.Subsequently, serial decimal dilutions of the cultures embedded in the filters were prepared in sterile saline solution, and transconjugants were selected by cultivation on LB agar containing appropriate antibiotics: cefotaxime (2 µg/mL); or rifampicin (100 µg/mL) and nalidixic acid (100 µg/mL) (Sigma-Aldrich, St. Louis, MO, USA).Donor and recipient strains were spread on LB agar supplement with cefotaxime, or with rifampicin and nalidixic acid, respectively, used as controls.All presumed transconjugants were confirmed to contain bla CMY-2 by PCR using previously described primers and conditions [74].Each experiment consisted of three biological replicates and three technical replicates.

Genome Annotation and Pangenome Analysis
All de novo assemblies of the isolates' genomes were carried out using the SPAdes (v.3.11.1), and the annotation was performed using the BAKTA annotation pipeline v1.7.0 with the default setting [75].Core and accessory genome comparison analyses of the ESCresistant E. coli isolates were performed using the Roary pangenome pipeline v3.13.0 [76].Roary produces a core gene alignment result from gff3 files created by BAKTA annotation.The "core" genes were identified in the isolates using a 99% identity cut-off.A concatenated core gene alignment of all isolates' core genes was generated and attached as a supplementary file (Table S2).The combined core gene alignment was used to construct an approximately maximum-likelihood phylogenetic tree using SNP sites [77] and the tree with FastTree v2.1 [78].The gene presence/absence file obtained by the Roary pangenome annotation pipeline and the core gene phylogenetic tree were visualized using a web-based interactive visualization of the genome tool Phandango [79].

In Silico Analysis for Sequence Type and Resistance Genes
Sequencing data from the isolates, except for the isolates reported in the previous study [20], were analyzed using the web-based Center for Genomic Epidemiology (CGE) tools (https://cge.cbs.dtu.dk/services/,accessed on 15 May 2023).STs and resistance genes or chromosomal point mutations yielding resistance to specific antibiotics were confirmed using MLST 2.0 [80] and ResFinder 4.1 [81], respectively.

Isolate Phylogenetic Analysis
To detect/compare SNP (single nucleotide polymorphism) differences among ST429 and ST1286 isolates, CSIPhylogeny (https://cge.dtu.dk/services/CSIPhylogeny/,accessed on 10 June 2023) was used on the CGE server with default settings and ignoring heterozygous SNPs.ST429 and ST1286 input sequences were mapped to the earliest ESC-resistant E. coli genomes, i.e., PS15001 and SB17103 for ST429 and ST1286, respectively, as the reference strains to call SNPs and searched for the previously described nucleotide variations [82].The following criteria for high-quality SNP calling and filtering were chosen: (i) select a minimum depth of 10× at SNP positions; (ii) select the minimum relative depth of 10% at SNP positions; (iii) select a minimum distance of 10 bp between SNPs; (iv) select minimum SNP quality of 30; (v) select minimum read mapping quality of 25; and (vi) select a minimum Z-score of 1.96.Site validation for each SNP position was performed.For bootstrap, 1000 replicates were generated to construct the tree.Two SNP matrices were created in MS Excel for each pair of strains, and their phylogeny trees were visualized using iTOL (http://itol.embl.de/,accessed on 15 June 2023) [83].
To further compare the relationship among ST429 isolates worldwide, a search for E. coli ST429 by the Achtman 7-gene MLST in EnteroBase [84] provided 371 results up to 15 June 2023 (http://enterobase.warwick.ac.uk, accessed on 15 March 2023).Of these, 80 came with relevant metadata to be considered for further analysis and are provided in a supplementary file (Table S4).The metadata of interest were the source, sample name, year of isolation, and country of origin.To compare this study's 22 ST429 Danish poultry isolates and the 80 Enterobase ST429 isolates, a phylogenetic analysis using an ad hoc core genome multilocus sequence typing (cgMLST) scheme with 2513 genes [20] was performed using Ridom Seqshere+ (v5.0.1, Ridom GmbH, Munster, Germany) [85] and visualized with a minimum spanning tree.

Plasmid Sequencing
To obtain high-quality IncK-bla CMY-2 plasmid sequences, 28 out of the 46 ESC-resistant E. coli isolates were selected to represent different STs, sources, and countries, and subjected to long-read sequencing (Table 2).All isolates were grown on MacConkey agar (Oxoid, Basingstoke, UK) with 1 mg/L cefotaxime (Sigma-Aldrich, St. Louis, MO, USA) overnight at 37 • C and DNA was then extracted using Genomic-Tip G/500 kit (Qiagen, Hilden, Germany) following the manufacturer's protocol.Libraries were constructed with the 1 D Ligation Barcoding Kit (catalog no.SQK-RBK114.96,ONT, Oxford, UK) according to the manufacturer's protocol and were sequenced on a R10.4.1 flowcell (FLO-MIN114) used with a MinION Mk1B sequencing device and sequenced with MinKNOW software v4.5.5 for 20-24 h.Long reads in the fast5 format were base called, demultiplexed, and converted into fastq format using Guppy v6.4.4 (ONT).The adaptor sequences were removed using Porechop v0.2.2 [86].Hybrid assembly of long and short reads using Unicycler v0.4.0 [87] resulted in circular contigs of the plasmid.We mapped the short Illumina reads to the plasmid contigs and performed error correction using CLC Genomic Workbench v.11.0.1 (QIAGEN, Aarhus, Denmark) by calling variants based on the mapping.Manual correction of errors was needed since homopolymer areas in sequences were especially problematic for the MinION technology.

Conclusions
In this study, we investigated the persistence and dynamics of 46 ESC-resistant E. coli and their IncK-bla CMY-2 plasmids, previously isolated from poultry and humans.Our results revealed that different E. coli STs carried highly similar IncK-bla CMY-2 plasmids, with the most common ST429 being isolated both from poultry and a human blood infection.The presence of highly similar plasmids in different E. coli STs could be due to the persistence of IncK-bla CMY-2 .Furthermore, ST429 E. coli bla CMY-2 isolates were found to occur globally pointing toward a common ancestor that has spread between the various reservoirs.The distribution of IncK-bla CMY-2 plasmids also provided evidence for the worldwide spread of IncK-bla CMY-2 producing ESC-resistant E. coli isolates.Further surveillance of IncK-bla CMY-2 plasmids in different E. coli STs in poultry production chain and humans should be carried out in order to increase our understanding of the dynamics of these ESC-resistant E. coli.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antibiotics13040349/s1,Table S1: Distribution of minimum inhibitory concentration (MIC) against antibiotics among ESC-resistant E. coli isolates in the study; Table S2: Presence or absence of annotated genes in ESC-resistant E. coli isolates (n = 46) in the study; Table S3: Single nucleotide polymorphisms (SNPs) matrices for ESC-resistant E. coli belonging to (a) ST429 and (b) ST1286; Table S4: List of 80 E. coli ST429 with associated metadata retrieved from the EnteroBase database, as well as supplemental minimum spanning trees (a-d) of the 80 Enterobase isolates and 22 ST429 isolates from this study; Table S5: Annotations for two IncK-bla CMY-2 plasmids from our study (pSB17040 in ST429 and pSB18004 in ST1286); Table S6: List of IncK plasmids carrying bla CMY-2 in E. coli strains representing 53 plasmids from GenBank (NCBI) and 28 plasmids from the present study; Figure S1: Comparison of Danish IncK-bla CMY-2 plasmids from our study using BRIG; Figure S2: Comparison of Danish and international IncK-bla CMY-2 plasmids using BRIG.Figure S3a-c

Figure 1 .
Figure 1.Antimicrobial-resistance gene determinants were identified in the IncK-blaCMY-2-positive ESC-resistant E. coli (n = 46) using Resfinder 2.0 on the CGE tool.The heatmap shows the presence or absence of antimicrobial-resistance genes in each isolate.Rows and columns represent isolates and predicted antimicrobial-resistance genes, respectively.Colors indicate the predicted resistance phenotype to different classes of antibiotics for each isolate based on genotype.

Figure 1 .
Figure 1.Antimicrobial-resistance gene determinants were identified in the IncK-bla CMY-2 -positive ESC-resistant E. coli (n = 46) using Resfinder 2.0 on the CGE tool.The heatmap shows the presence or absence of antimicrobial-resistance genes in each isolate.Rows and columns represent isolates and predicted antimicrobial-resistance genes, respectively.Colors indicate the predicted resistance phenotype to different classes of antibiotics for each isolate based on genotype.

Figure 2 .
Figure 2. Pangenome analysis of all ESC-resistant E. coli carrying IncK-bla CMY-2 (n = 46).(a) Distribution of total genes (100%): core genes (27.91%) found in ≥99%, soft-core genes (3.17%) found in between 95% and 99%, shell genes (21.10%) found in between 15% and 95%, and cloud genes found in <15% of the E. coli in the study.(b) Maximum likelihood phylogenetic tree inferred from the alignment of the 3094 core genes of 46 ESC-resistant E. coli by FastTree.(c) Annotation of gene presence (blue) and absence (white) matrix across the pangenome of the E. coli.The top scale shows the complete genome size (kbp).Each row shows the gene content of an E. coli isolate.Each column shows the comparative gene clusters.The data were visualized using Phandango.

Figure 3 .
Figure 3. Phylogenetic trees show the relationships between ST429 (a) and ST1286 (b) ESC-resistant E. coli isolates collected from Danish, Finnish, and German poultry and one Danish human sample (see Table2for isolate information).Each SNP-based tree was constructed with CSI Phylogeny 1.4 (https://cge.dtu.dk/services/CSIPhylogeny/,accessed on 15 May 2023), using the genomes of PS15001 and SB17103 (in red) as a reference in each case.For ST429, poultry isolates from Finland and Germany and 1 human isolate from Denmark are shown in black bold.Isolates from the same date and source are presented in the same color.The observed number of SNPs among these isolates is indicated as well.

Figure 4 .
Figure 4. Minimum spanning tree of 102 ST429 E. coli isolates collected from poultry production (this study, Danish isolates, n = 22) and Enterobase (n = 80) based on an ad hoc 2513 gene cgMLST scheme and 7 E. coli MLST Warwick targets calculated in Ridom Seqshere+.Isolates are colored according to country and year of isolation indicated in circles (each circle may contain more than one isolate in cases of no allelic differences).The number of isolates per country is indicated in brackets in the color legend.Numbers of allelic differences between isolates are shown in the connecting lines.More detailed information on isolates and the minimum spanning trees can be found in Supplementary TableS4and FigureS3a-d.

Figure 6 .
Figure 6.FastTree approximate maximum-likelihood phylogenetic tree of IncK-blaCMY-2 plasmids constructed with plasmids from this study (n = 28) labeled in purple and from GenBank, NCBI, labeled in black font (n = 53).The heatmap denotes the presence/absence of antimicrobial-resistance and virulence genes on the plasmids.The following annotation from left to right represents country, species, source, and year of isolation.

Figure 6 .
Figure 6.FastTree approximate maximum-likelihood phylogenetic tree of IncK-bla CMY-2 plasmids constructed with plasmids from this study (n = 28) labeled in purple and from GenBank, NCBI, labeled in black font (n = 53).The heatmap denotes the presence/absence of antimicrobial-resistance and virulence genes on the plasmids.The following annotation from left to right represents country, species, source, and year of isolation.

Table 1 .
Distribution of MICs of several antibiotics in the ESC-resistant E. coli (n = 46) isolates included in the study.
* Black vertical lines represent epidemiological cut-off values (ECOFFs) for resistance.Light blue fields represent the range of dilutions tested for each antimicrobial agent, and light gray fields represent the non-tested dilutions.

Table 2 .
Overview of characteristics of IncK-bla CMY-2 plasmids present in the 46 ESC-resistant E. coli isolated from poultry, meat, and clinical samples, collected in different countries over the period 2012-2020.