Genomic Features of MCR-1 and Extended-Spectrum β-Lactamase-Producing Enterobacterales from Retail Raw Chicken in Egypt

Colistin is considered as a last resort agent for treatment of severe infections caused by carbapenem-resistant Enterobacterales (CRE). Recently, plasmid-mediated colistin resistance genes (mcr type) have been reported, mainly corresponding to mcr-1 producers. Those mcr-1-positive Enterobacterales have been identified not only from human isolates, but also from food samples, from animal specimens and from environmental samples in various parts of the world. Our study focused on the occurrence and characterization of mcr-1-positive Enterobacterales recovered from retail raw chicken in Egypt. From the 345 retail chicken carcasses collected, a total of 20 samples allowed to recover mcr-1-positive isolates (Escherichia coli, n = 19; Citrobacter freundii, n = 1). No mcr-2- to mcr-10-positive isolate was identified from those samples. The colistin resistance trait was confirmed for all those 20 isolates with a positivity of the Rapid Polymyxin NP (Nordmann-Poirel) test. Minimum inhibitory concentrations (MICs) of colistin for all MCR-1-producing isolates ranged between 4 and 16 μg/mL. Noticeably, 9 out of the 20 mcr-1-positive isolates produced an extended-spectrum β-lactamase (ESBL), respectively producing CTX-M-9 (n = 2), CTX-M-14 (n = 4), CTX-M-15 (n = 2), and SHV-12 (n = 1). Noteworthy, the fosA4 gene encoding resistance to fosfomycin was found in a single mcr-1-positive E. coli isolate, in which both genes were located on different conjugative plasmids. The pulsed-field gel electrophoresis (PFGE) patterns were identified, corresponding to 10 different sequence types (STs), highlighting the genetic diversity of those different E. coli. Whole-genome sequencing revealed three major types of mcr-1-bearing plasmids, corresponding to IncI2, IncX4, and IncHI2 scaffolds. The occurrence of MCR-1-producing multidrug-resistant Enterobacterales in retail raw chicken is of great concern, considering the possibility of transmission to humans through the food chain.

The mcr-1 gene has been reported mostly from Enterobacterales recovered from food, animals, humans, and the environment in various parts of the world [19]. The extensive use of colistin in veterinary medicine for prophylactic and therapeutic purposes as well as for growth promotion has been recognized as a major risk factor for the emergence and dissemination of colistin-resistant Escherichia coli isolates in food-producing animals [19,20]. Hence, the food chain may significantly increase the dissemination and acquisition of colistin-resistance worldwide.
The mcr-1-carrying plasmids in E. coli identified from human and from poultry meat in Switzerland were highly similar, indicating that certain types of epidemic plasmids such as IncI2 and IncX4 play an important role in the dissemination of the mcr-1 gene along the food chain and in humans [20]. In Africa, mcr-1-positive E. coli have been detected in food items, including chicken meat and chicken carcasses [21,22], sausage [23], cheese [24], and raw milk [25]. However, there is still a lack of comprehensive data concerning the epidemiology of mcr genes in E. coli of food origin in African countries, including Egypt [26]. This study reports the isolation and characterization of foodborne E. coli strains carrying mcr-1 gene, fosA, and extended-spectrum ß-lactamase (ESBL)-encoding genes recovered from retail raw chicken in Egypt.

Bacterial Isolates, Susceptibility Testing
Between July and December 2018, a total of 345 retail chicken carcasses were randomly collected from different poultry slaughterhouses, supermarkets, and butcher shops in different Egyptian cities (Qena, Luxor, Nag Hammadi, and Esna city). The neck skin samples (25 g) taken from each carcass were homogenized and enriched in buffer peptone water (225 mL) for 24 h at 37 • C with shaking. Colistin-resistant isolates were recovered by direct spreading on MacConkey agar supplemented with 2 µg/mL colistin. Colonies of different morphology, size, and color from each plate were selected and used for further analysis, and subsequently submitted to a screening of mcr genes. For that purpose, DNA was extracted from all colistin-resistant strains with the QIAamp DNA mini-kit and the QIAcube workstation (Qiagen, Courtaboeuf, France), according to the manufacturer's instructions. Then, PCR screening for plasmid-mediated colistin resistance genes (mcr-1 to 10) was performed [12,27]. mcr-positive isolates were subsequently selected for further characterization and analysis. Antimicrobial susceptibility testing was performed by the disk diffusion method according to the Clinical and Laboratory Standards Institute (CLSI) recommendations (CLSI M100 ED30:2020), on Muller-Hinton agar plates broad-spectrum cephalosporins, carbapenems, aminoglycosides, aztreonam, quinolones, sulfonamides, tigecycline, and fosofmycin. Minimum inhibitory concentrations (MICs) of colistin were evaluated by broth microdilution in cation-adjusted Mueller-Hinton broth (Bio-Rad, Cressier, Switzerland), as recommended by the European Committee on Antimicrobial Susceptibility Testing (EUCAST)/Clinical & Laboratory Standards Institute (CLSI) joint guidelines (https://eucast.org/fileadmin/src/media/PDFs/EUCAST_files/General_ documents/Recommendations_for_MIC_determination_of_colistin_March_2016.pdf). Results were interpreted according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST)/CLSI joint guidelines (www.eucast.org). The MIC of fosfomycin was determined by the agar dilution method using cation-adjusted Mueller-Hinton agar (MHA-CA, reference 64884; Bio-Rad, Marnes-la-Coquette, France) supplemented with glucose-6-phosphate (25 µg/mL), as recommended by CLSI, 2020 (CLSI M100 ED30:2020).

Phenotypic Confirmation of Resistance Patterns
Extended-Spectrum β-Lactamase (ESBL)-producing and colistin-resistant isolates were tested by the Rapid ESBL NP (Nordmann-Poirel) test and the Rapid Polymyxin NP test respectively [28,29], in order to confirm their resistance patterns. The Rapid Fosfomycin NP test was performed as described [30]. For all selected isolates, identification at the species level was performed using the API20E system (bioMérieux, La-Balme-les-Grottes, France).

Molecular Assays for Other Resistance Determinants
The identification of ESBL-encoding genes (namely bla TEM , bla SHV , and bla CTX-M ) was performed by PCR using previously reported primers [31]. Additionally, PCR amplification was performed to detect any known plasmid-mediated fosA genes (fosA1 to fosA8), as previously described [32]. All positive PCR amplicons were sent for sequencing (Microsynth, Balgach, Switzerland).

Mating-Out Assays
The transferability of all mcr-1-positive isolates was investigated by a filter-mating assay, as described previously [33], in which mcr-1-positive isolates were used as donors and azide-resistant E. coli J53 as the recipient. Transconjugants were selected on LB agar supplemented with sodium azide (100 µg/mL) and colistin (2 µg/mL). Transconjugants were obtained for all donors at 25-30 • C or 37 • C. Transconjugants were confirmed by PCR targeting the mcr genes, and antimicrobial susceptibility testing.

Plasmid Isolation, Plasmid Analysis
The plasmids carrying the mcr-1 gene from E. coli transconjugants were typed by using the PCR-based replicon typing (PBRT) method, as previously reported [34]. The PCR scheme was complemented with primers specific for the IncX3-type plasmids [35], as well as primers specific for IncX4 plasmids [36]. The size of the plasmid was obtained after Kieser extraction for the resulting E. coli transconjugant strains [37], followed by gel electrophoresis analysis, with reference strain E. coli 50,192 containing 4 plasmids (154, 66, 48, and 7 kb, respectively) being used as a molecular marker.

Plasmid Sequencing and Bioinformatic Analysis
Seven representative mcr-1-plasmids were selected for sequencing by using Illumina technology. Templates used corresponded to plasmid extracts obtained for the E. coli transconjugants using the Qiagen Large Plasmid Construct kit (Qiagen, Hilden, Germany). Genomic libraries were assessed using a Nextera XT library preparation kit (Illumina Inc., San Diego, CA, USA), and sequencing was performed using an Illumina MiniSeq system with 150 bp paired-end reads. De-novo genome assembly was performed using the CLC Genomic Workbench (version 20.0.4; CLC Bio, Aarhus, Denmark), and contigs with a minimum contig length of 800 nt were generated. The resulting assembled sequences were uploaded to the Center for Genomic Epidemiology server (http://www.genomicepidemiology.org/). The MLST and plasmid replicon types were determined using the MLST (version 2.0.4) and PlasmidFinder (version 2.0.1) software, respectively. Antimicrobial resistance was analyzed by ResFinder 4.1 (https://cge.cbs. dtu.dk/services/ResFinder/) and CARD (https://card.mcmaster.ca/analyze/rgi). The constructed plasmids were annotated automatically by the RAST server using the RAST-tk scheme [39], followed by manual inspection and correction using the BLASTn and BLASTp tools (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The circular image and comparisons between other reported similar plasmids were performed using the BLAST Ring Image Generator (BRIG) tool ( Figure 1) [40].
Microorganisms 2021, 9, x FOR PEER REVIEW 4 of 12 ing the CLC Genomic Workbench (version 20.0.4; CLC Bio, Aarhus, Denmark), and contigs with a minimum contig length of 800 nt were generated. The resulting assembled sequences were uploaded to the Center for Genomic Epidemiology server (http://www.genomicepidemiology.org/). The MLST and plasmid replicon types were determined using the MLST (version 2.0.4) and PlasmidFinder (version 2.0.1) software, respectively. Antimicrobial resistance was analyzed by ResFinder 4.1 (https://cge.cbs.dtu.dk/services/ResFinder/) and CARD (https://card.mcmaster.ca/analyze/rgi). The constructed plasmids were annotated automatically by the RAST server using the RAST-tk scheme [39], followed by manual inspection and correction using the BLASTn and BLASTp tools (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The circular image and comparisons between other reported similar plasmids were performed using the BLAST Ring Image Generator (BRIG) tool ( Figure 1) [40].

Results and Discussion
Occurrence of MCR-1 producers: From the 345 retail chicken carcasses, a total of twenty samples allowed to recover mcr-1-positive isolates (E. coli, n = 19; Citrobacter freundii, n = 1). No mcr-2to mcr-10-positive isolate was identified from those samples. The colistin resistance trait was confirmed for those twenty isolates by the positivity of the Rapid Polymyxin NP test. Determination of the MICs of colistin for all MCR-1-producing isolates showed that they ranged from 4 to 16 µg/mL (Table 1). Although the current situation of antimicrobial use for livestock is not precisely known in Egypt, colistin usage is common in animal husbandry in Egyptian farms, including for calves, poultry (for example, treatment of colibacillosis), and rabbits [41,42]. Therefore, it is tempting to speculate that such selection of colistin-resistant E. coli isolates among Egyptian farms might be related to the overuse of colistin in the local farming industry [41]. Two mcr-1-positive E. coli strains isolated from humans have been identified in Egypt so far [43]. The mcr-1 gene had also been identified from E. coli isolates recovered from a diseased cow [44], and from healthy broilers [45]. Concerning strains isolated from food, a single mcr-1-positive E.
coli strain was isolated from cheese [24] and beef sausage [23] in Egypt.
Identification of FosA4 fosfomycin resistance determinant: It is noteworthy that, among these twenty mcr-1-positive isolates identified in this study, a single isolate was found to be fosfomycin-resistant (positive Rapid Fosfomycin NP test, MIC of fosfomycin, >256 mg/L) that was shown to be associated to the presence of the fosA4 gene, which had not been co-transferred with mcr-1. In that isolate, the mcr-1 and fosA4 genes were located on IncX4 and IncFII plasmids, respectively. The fosA4 gene was previously reported in the literature from human clinical E. coli isolates in Japan and Australia [56,57]. The identification of FosA determinants remains rare among animal isolates, with few reports so far, corresponding to the occurrence of the fosA3 gene among E. coli isolates from different animal sources (cattle, pigs, poultry, and pet animals) in China, Brazil, and France [58][59][60][61][62]. To our knowledge, our study reports the first fosA4-positive E. coli isolate from an animal source. Plasmid characterization: Conjugative assays and PCR-based replicon typing (PBRT) analysis were performed for all mcr-1-positive isolates. Transconjugants were obtained for all donors, indicating that the mcr-1 gene was located on self-transferable conjugative plasmids (Table 1). Those E. coli transconjugants exhibited MIC values for colistin at 4 or 8 µg/mL, while that of the E. coli J53 recipient strain was found at 0.25 µg/mL. No additional resistance trait was transferred along with the mcr-1 gene in any of the E. coli transconjugants, except for EC65.2 TC that showed resistance to tetracycline, chloramphenicol, and sulfonamides. PBRT analysis revealed that the mcr-1 gene was localized on different plasmid scaffolds differing in sizes and structures, including IncI2 (55%), IncX4 (30%), and IncHI2 (15%) ( Table 1). Those findings are consistent with previous reports that the worldwide spread of the mcr-1 gene is mainly driven by three major plasmid types-IncI2, IncX4, and IncHI2 [1,19,[63][64][65][66][67][68]-with IncI2 being the most prevalent plasmid backbone, followed by IncX4 and IncHI2.
Considering that several plasmid types were identified, seven representative mcr-1positive plasmids were entirely sequenced. WGS data analysis showed that the seven plasmids belonged to three types of plasmids, including IncX4 of size ranging from 31.8 to 32.1 kb (n = 3), IncI2 of size ranging 64.2-67.7 kb (n = 3), and IncHI2 of 195.8 kb in size (n = 1). None of the mcr-1-positive IncI2 and IncX4 plasmids possessed additional resistance determinants, confirming the phenotypic observations made with the E. coli transconjugants, whereas multiple resistance elements were detected alongside mcr-1 in the pEGYMCR65 plasmid. The three sequenced IncX4 type plasmids (pEGYMCR8, pEGYMCR16, pEGYMCR60) were nearly identical, and showed typical plasmid backbones encompassing genes encoding proteins involved in replication, maintenance, and transfer.
Interestingly, detailed sequence analysis of the IncX4 plasmids showed that they were almost identical to the mcr-1-positive plasmid pCFSAN061769_01 (~97% query coverage and~99.9% sequence identity; GenBank accession no. CP042970.1) identified in an E. coli recovered from cheese in Egypt, indicating that this plasmid type is circulating among different sources. The sequence of those IncI2 plasmids (pEGYMCR17, pEGYMCR23, pEGYMCR62) was almost identical to plasmid pEGYMCR-1, identified in a single E. coli isolate recovered from a meat product (beef sausage) in Egypt, and to plasmid pMCR-GN775 (accession no. KY471307) identified in an E. coli strain recovered from a gastrostomy tube site and rectum of a patient hospitalized in Canada who noticeably received previous healthcare in Egypt (Figure 1) [23,69]. This basically underlines that the dissemination of the mcr-1 gene is at least partially linked to that of "epidemic" plasmids.
The sequence of plasmid pEGYMCR65 was highly similar to that of the mcr-1-positive IncH12 plasmid, pCFS3292-1 (95% query coverage and 99.99% sequence identity; GenBank accession no. CP026936.2), identified from an E. coli of bovine origin recovered from animals presenting with diarrhea and mastitis in France. Also, plasmid pEGYMCR65 was highly similar to the mcr-1-positive plasmid pEGY1-MCR-1 (91% query coverage and 99.99% sequence identity; NCBI Reference Sequence: NZ_CP023143.1) identified in E. coli isolated from cheese in Egypt (Figure 1), that latter missing transposon Tn21 encoding mercury resistance genes (mer operon).
Genetic environment of the mcr-1 gene: In a recent study, an intermediate circular form of the insertion sequence ISApl1 associated with mcr-1 was detected, suggesting that ISApl1 might play a driving role in the horizontal gene transfer of this resistance gene [70]. ISApl1 (usually only one copy) is found upstream of the mcr-1 gene. Recently, it has been reported that a second copy of ISApl1 may be found downstream of the mcr-1 gene, therefore forming the composite transposon Tn6330 [71,72]. By further analysis of the genetic context surrounding the mcr-1 genes identified in this study, the mcr-1-pap2 element was identified on all plasmids. The intact composite transposon Tn6330 (ISApI1-mcr-1-orf -ISApI1) was identified on IncHI2 plasmids. IncX4 and IncI2 plasmids did not display an ISApl1 element in the vicinity of the mcr-1 gene (Figure 1). Our results are consistent with previous reports showing that circular intermediates could be detected mainly in IncHI2 plasmids, not in IncX4 or IncI2 plasmids [23,54,69,70,73].

Conclusions
Our findings highlight the occurrence of MCR-1-and ESBL-producing Enterobacterale strains in poultry farms in Egypt. It might be speculated that the overuse and misuse of both β-lactams and colistin in veterinary practice might be the main cause of such worrying phenomenon. A high diversity of genetic structures responsible for the acquisition of mcr-1 in Enterobacterales, and particularly in E. coli, was shown here by focusing on those isolates recovered from retail chicken meat from Egypt. The detection of plasmid-mediated colistin resistance in poultry meat is of potential public health concern since it is considered as an important source of transmission of plasmid-mediated mcr-1 to consumers through handling and eating of raw or undercooked meat. Our findings suggest that the local and global spread of the mcr-1 gene has mainly been driven by three major plasmid types, with IncI2 being the most prevalent plasmid backbone, followed by IncX4 and IncHI2. Effective intervention approaches aiming to reduce antibiotic selective pressure in food-producing animals in low-income countries including Egypt must absolutely be implemented, to prevent further selection of multidrug-resistant bacteria.

Data Availability Statement:
The complete nucleotide sequences of pEGYMCR_IncHI2, pEGYMC R_IncI2, and pEGYMCR_IncX4 were deposited as GenBank accession numbers MT499884, MT499885, and MT499886, respectively.

Conflicts of Interest:
The authors declare no conflict of interest.