Prevalence and characterisation of antimicrobial resistance genes and class 1 and 2 integrons in multiresistant Escherichia coli isolated from poultry production

A global increase in the populations of drug resistant bacteria exerts negative effects on animal production and human health. Our study has been focused on the assessment of resistance determinants in relation to phenotypic resistance of the 74 commensal E. coli isolates present in different ecological environments. The samples were collected from poultry litter, feces, and neck skin. Among the microorganisms isolated from the poultry litter (group A), the highest resistance was noted against AMP and DOX (100%). In the E. coli extracts from the cloacal swabs (group B), the highest resistance was observed against AMP (100%) and CIP (92%). The meat samples (group C) were characterized by resistance to AMP (100%) and STX (94.7%). Genes encoding resistance to β-lactams (blaTEM, blaCTX-M), fluoroquinolones (qnrA, qnrB, qnrS), aminoglycosides (strA-strB, aphA1, aac(3)-II), sulfonamides (sul1, sul2, sul3), trimethoprim (dfr1, dfr5, dfr7/17) and tetracyclines (tetA, tetB) were detected in the studied bacterial isolates. The presence of class 1 and 2 integrons was confirmed in 75% of the MDR E. coli isolates (plasmid DNA), of which 60% contained class 1 integrons, 15% contained class 2 integrons, and 11.7% carried integrons of both classes. Thus, it may be concluded that integrons are the common mediators of antimicrobial resistance among commensal multidrug resistant Escherichia coli at important stages of poultry production.

The increasing resistance to commonly applied antimicrobial agents is being reflected by growing multiple drug resistance (MDR) in bacteria and is becoming a growing threat to public health. The use of antimicrobial agents in animal husbandry has been linked to the development and spread of the resistant bacteria 1 . Resistant bacteria can be transferred for example from poultry products to humans via consuming or handling meat contaminated with pathogens 2 . However, the resistance of commensal bacteria is equally important as they constitute a reservoir and vector of resistance determinants in the environment 3 .
Exposure to antimicrobials of different classes can lead to cross-resistance and the selection of antibiotic resistance genes (ARGs) that may spread laterally on mobile genetic elements (MGEs) via horizontal gene transfer (HGT) 4 . HGT is a phenomenon in which genes are transferred between organisms of either the same or different species which often remain in a close ecological relationship 5 .
It has been shown that conjugation is one of the key mechanisms responsible for the spread of the ARGs 6 . One of the most efficient mechanism of acquiring ARGs is facilitated by integrons-a site-specific recombination systems capable of recruiting open reading frames (ORF) in the form of mobile gene cassettes 7 . Integrons are divided into two distinct subsets, mobile integrons (MIs)-linked to mobile DNA elements, which are primarily involved in the spread of ARGs, and chromosomal integrons (CIs). MIs are associated with conjugation plasmids

Results
Antimicrobial resistance phenotypes of E. coli isolates. We have analyzed the incidence of multidrug resistance gene sequences and the prevalence of class 1 and 2 integrons within 74 commensal E. coli isolates, obtained from poultry litter (group A, n = 23), swabs from broiler chicken cloaca (group B, n = 26) and poultry meat (group C, n = 25). 60 (81.1%) of them exhibited a multiresistant phenotype (resistance to at least three different antimicrobial agent families). In the first step of study, the phenotypic resistance of the E. coli isolates to six antibiotics and chemotherapeutics was assessed.
Out of 23 E. coli isolates obtained from poultry litter, 16 showed multidrug resistance. The highest resistance was recorded for AMP (100%), DOX (100%), CIP and STX (81.3%), and AMC (75%); the lowest for GEN (12.5%). Out of 26 isolates obtained from chicken cloaca, 25 exhibited MDR. Among the examined MDR isolates, the highest percentage of resistance was observed for the following antibiotics: AMP (100%), CIP (92%), AMC, DOX (88%), STX (84%) and the lowest for GEN (36%). All E. coli isolates obtained from the cloaca of chickens, showed phenotypic resistance to the antibiotics classes with which the broilers were treated on farms. Among E. coli isolates obtained from meat, 19 of them showed MDR. In this group, the highest resistance was observed for AMP (100%). Lower resistance was noted for: STX (94.7%), CIP (78.9%), DOX (68.4%). Resistance to AMC and GEN was noted in 42.1% and 36.8% of E. coli isolates respectively. Figure 1., shows the percentage of phenotypic resistance to 6 antibiotics among MDR E. coli isolates obtained from poultry litter, cloacal swabs and poultry meat (group A, B and C). Antimicrobial susceptibility tests showed that 11.7% of all MDR isolates of E. coli, were resistant to three antibiotics, 38.3% to four, and 35% showed resistance to five and 15% to six antibiotics (Table 1). Among all E. coli isolates positive for integron sequences, the most common drug resistance profile was that of the resistance to: AMP, AMC, CIP, STX, DOX. In the case of 2 isolates with a pair of integrons, E. coli isolates that showed resistance to 6 antimicrobial agents were reported. Their resistance profile was the same: AMP, AMC, CIP, GEN, STX, DOX. Identification and characteristics of resistance genes.. Resistance to AMP and AMC encoded by the narrow spectrum beta lactamase resistance gene (bla TEM ) was found in genomic DNA of all E. coli isolates from groups A and B, and in 63,2% of the poultry meat (group C). In case of other genes encoding beta-lactamase resistance, the occurrence of bla CTX-M gene was noted in one colon isolate. Bla SHV gene was detected in one isolate from the poultry meat swabs. Among the MDR isolates showing the ciprofloxacin (CIP) resistance phenotype, a total of 7 qnrA genes, 10 qnrB genes and 6 qnrS genes were reported in genomic DNA. 18 MDR isolates were gentamicin-resistant, and the strA-strB, aphA1, and aac(3)-II genes, giving resistance to aminoglycosides, were present in: 13 isolates of E. coli obtained from litter, 19 isolates from cloacal swabs and 5 isolates from poultry meat. 13 of the 16 tested E. coli MDR isolates showed a sulphonamides resistance phenotype which was encoded by: sul1 (50%), sul2 (18.8%) and sul3 (25%). In the case of 21 MDR isolates obtained from cloaca the sul1 and sul3 genes were recorded in 28% and the sul2 gene in 44% of the cases. In this group, the presence of pairs of sul genes was noted in 5 isolates, in the su1-sul2 combination. In the MDR group of meat isolates, 18 E. coli isolates contained the following genes: sul1 (31.6%), sul2 (26.3%) and sul3 (10.5%) and in one case a pair of genes (sul1 and sul2) were noted. In the case of E. coli isolates recovered from the litter, one of them confirmed 3 genes (sul1, sul2, sul3) determining resistance to sulfonamides. Sulfonamide-resistant isolates of E. coli showed in most cases The tetC gene was not found in any of the studied groups. The pairs of tetA and tetB genes were found in groups A and B in 4 and 3 cases, respectively. The prevalence of the resistance genes among multidrug resistant E. coli isolates obtained from poultry litter (Group A), cloacal swabs (Group B) and poultry meat (Group C) is shown in Table 2 and   www.nature.com/scientificreports/ much smaller number of cases, only in cloacal and meat isolates (4 and 5 cases). The frequency of the integtons of class 2 differed significantly (P ≤ 0.05) between these two locations. In the samples in which both classes of the integrons were detected (group B and C), the class 1 integrons were significantly more frequent than class 2 integrons (B: P ≤ 0.01, and C: P ≤ 0.05).
In the group of MDR E. coli isolated from the poultry meat (group C), genes contained in the class 1 integron variable region were detected in 9 cases (Table 5). VR class 1 integrons showed less variation compared to group B isolates and usually contained 2 gene cassette arrays: aadA1 (22.2%) and dfrA1-aadA1 (77.8%). The variable region of 5 class 2 integrons contained in only two cases a set of cassettes: dfrA1-sat2-aadA1. In one strain, a pair of class 1 and 2 integrons was detected as well as their variable parts were empty.
In our study, 74 unrelated commensal isolates of Escherichia coli originated from poultry litter, cloacal swabs and poultry meat were phenotypically and genotypically tested for the antimicrobial resistance and the presence of integrons as factor, for the development of antibiotic resistance and the emergence of MDR strains. Among them, 60 isolates (81.1%) were multiresistant (resistant to a minimum of three classes of antibiotics). We have found the highest level of antimicrobial resistance (96.2%) in the E. coli isolates obtained from broiler intestinal swabs (group B).
The high incidence of multidrug resistance in our study, particularly regarding isolates obtained from feces and meat, is extremely significant and should be regarded as a serious health risk due to the fact, that multidrug resistant isolates may have a chance of contaminating food products, and consequently being transferred to humans.
The percentage of resistance to some antimicrobial agents (Ampicillin, Doxycycline, Trimetophrim, Sulfamethoxazole, and Ciprofloxacin) in all studied groups was particularly high (100-68.4%), which indicates that the commensal E. coli isolates may be a reservoir of resistance to antibiotics and chemotherapeutics. Our data largely overlaps with the data made available by the European Food Safety Authority and the European Centre for Disease Prevention and Control, on the resistance profile of the commensal E. coli isolates obtained from slaughterhouse broilers, collected between 2009 and 2014 in Poland 28 . It confirms high resistance of E. coli isolates to nalidixic acid ciprofloxacin, and ampicillin (70-90%) and a limited resistance to tetracyclines, sulfonamides and streptomycin 28,29 .  n  25  25  22  23  9  21  22  25  1  0  0  1  4  3  14  1  4  7  11  7  15  0  4  13  9  0   %  100  88  92  36  84  88  100  4  0  0  4  16  12  56  4  16  28  44  28  60  0  16  52  36  0   Group  C   n  19  19  8  15  7  18  13  12  0  0  1  5  2  1  3  2  0  6  5  2  5  2  0  6  The bla TEM gene encoding β lactamase, which gives resistance to penicillins and cephalosporins of the first generation, has been detected in all multidrug-resistant isolates obtained from litter and feces. The dominance Quinolone resistance is a current worldwide problem in human and veterinary medicine 33 . Quinolone resistance can encoded in bacterial chromosome or be present in plasmids. Plasmid-mediated quinolone resistance (PMQR) promote the spread of the multi-drug resistance phenotype. For example, qnr genes present on MDR plasmids are often found with genes encoding β-lactamases 34 . Of all Qnr determinants present in the our study, the qnrB gene was found most frequently. Similar results were published in other studies 33,35 . The occurrence of PMQR is also associated with resistance to other groups of antibiotics. The mechanism responsible for this phenomenon is related to the presence of the aminoglycoside acetyltransferase enzyme-AAC(6')-Ib-cr, which modifies both the molecular structure of some fluoroquinolones and aminoglycosides or the oqxAB gene encoding an MDR-type efflux pump contributing to increased resistance to quinolones and chloramphenicol, trimethoprim and quinolones 36,37 . In Seo and Lee 33 study, 10 PMQR-positive E. coli were isolated from chicken meat, and these isolates also showed higher resistance rates to several antimicrobial agents when compared to PMQR-negative E. coli. This is consistent with previous studies showing that the PMQR genes increase resistance to other antimicrobials and cause MDR to drugs such as: aminoglycosides, β-lactams, chloramphenicol, sulfonamides, tetracyclines and trimethoprim 38 .
Tetracyclines are commonly used to treat bacterial infections in livestock, including poultry in many countries [39][40][41] . Due to the numerous advantages of tetracyclines, such as their widespread availability, low cost, and several side effects, the use of such antibiotics to treat animal and human infections has been increasing in recent years 42 . The chickens are treated with tetracycline orally and their metabolites (up to 90%) are excreted in the feces 43 on manure 44 . It is noteworthy that in our study the highest resistance to doxycycline was recorded among E. coli isolates derived from poultry litter, which is a mixture of poultry manure, litter, feathers, feed, and spilled drinking water that accumulates during breeding 45 . However, the proportion of E. coli isolates with resistance to tetracyclines was lower than the proportion of E. coli isolates resistant to beta lactam antibiotics. Similar results were obtained in study by Islam et al. 46 in which MDR isolates were the most resistant to tetracyclines (96.6%) and penicillins (100%). In addition to the antibiotic residues, manure also contains MDR bacteria and resistance genes, which can be transmitted to humans through direct contact between poultry and humans or indirectly via the food chain 47 . The results of genotyping showed that similarly to other published data, the resistance of commensal E. coli to tetracyclines was induced by the presence of tetA and tetB genes 48 . The highest content of tet genes in poultry litter isolates confirms the thesis put forward by Furtula et al. 44 , that the breeding environment significantly contributes to the spread of the resistance, via the transmission of the resistance genes.
Resistance to sulfonamides in Gram-negative bacteria generally results from the presence of the genes sul1, sul2, and/or sul3. Among them, the sul2 gene is the most widely distributed sul gene in porcine, avian, or human E. coli isolates, and it plays an important role in sulfonamide resistance 49 . In our research, the prevalence of sul2 Table 3. Phenotypes, antibiotic resistance and prevalence of class 1 and 2 integrons and their resistance gene cassettes in MDR E. coli isolates obtained from poultry litter. (Group A). www.nature.com/scientificreports/ genes was highest in isolates from cloaca swabs (44%) and was similar to the results of other studies 50,51 . Interestingly, in only one strain obtained from litter, all tested sul genes were determined. The frequency of sul genes detection in our experiment corresponded to other studies 50,51 . The selective pressure exerted by sulfonamides in the poultry industry appears to be high, which may favor the maintenance of acquired sul genes among bacteria 52 . In our study, the total prevalence of integrons (75%) in MDR isolates was higher than the prevalence of integrons detected in other poultry production prevalence studies 53,54 . We also found a clear predominance of class 1 integrons in relation to class 2 integrons, which is consistent with previous studies that also showed the highest prevalence of this class in poultry isolates 55,56 .

Detection of intI1 Genes included in cassettes Detection of intI2 Genes included in cassettes
Most of the integrons detected in our study contained gene cassettes encoding resistance to trimethoprim (dfrA gene type) and streptomycin/spectinomycin belonging to aminoglycosides (aadA gene type), and the most frequently detected sequence of cassettes was dfrA1-aadA1. The persistence of these genes, which have been reported worldwide in isolates from different sources, may be related to the widespread use of streptomycin/ spectinomycin, trimethoprim, sulfonamides and other antibiotics in food producing animals. Although the afore-mentioned aminoglycosides are not used therapeutically in animals in Poland, the presence of aadA genes may be a form of genetic memory, in case of re-exposure of the microorganism to this group of antibiotics 57 .
The analysis of the variable part of integrons in our experiments indicated the presence of one to three gene cassettes. In group B, we noted the highest number of integrons among all tested groups and a greater variety of gene cassettes within the integron variable part. The higher prevalence of class 1 integrons among the E. coli isolates obtained from the cloaca may be caused by the development and spread of the resistance genes, due to the misuse or abuse of antibiotics in the poultry production 1 .
Of all MDR E. coli isolates that had integrons, 8 isolates (13.3%) did not contain any of the evaluated gene cassettes. The situation regarding the so-called "empty integrons" has already been described by Fonseca et al. 58 , where it was indicated that these bacteria could rapidly develop into MDR in the future. However, it cannot be ruled out that integrons may have previously removed cassettes of resistance genes acquired by cutting them out for unknown reasons 59 .
The data obtained in this study highlights the importance of commensal E. coli in the spread of resistance genes at different stages of poultry production. We confirmed that more than half of the multidrug-resistant Table 4. Phenotypes, antibiotic resistance and prevalence of class 1 and 2 integrons, detected in 25 MDR E. coli isolates obtained from cloacal swabs (Group B). www.nature.com/scientificreports/ isolates (75%) contained integrons. Furthermore, we showed that antibiotic resistance can also occur on nonintegron structures. Therefore, there is a need for further detailed genetic studies on the evolution of isolates present in poultry to uncover the underlying mechanisms the acquisition of resistance by these microorganisms and to analyze the implications for humans. Such data may be used to determine the dynamics of resistance development and strategies to counteract antibiotic resistance among zoonotic microorganisms transmitted through food of animal origin at all stages of the food chain, from farm to table.

Detection of intI1 Genes included in cassettes Detection of intI2 Genes included in cassettes
Materials and methods E. coli isolates. A total number of 74 E. coli isolates was collected from three areas of poultry production: litter swabs from chicken houses (n = 23), cloacal swabs from chicken (n = 26), chicken meat from slaughterhouses (n = 25). All samples were collected between November 2019 and March 2020. Litter samples from 23 different chicken houses were acquired in accordance with the boot swabs sampling procedure guidelines of the national control program for Salmonella serotypes in poultry flocks in line with the guidelines of the current EU law 60 . Samples were collected from 4-week-old chicks (average weight 1.6 kg) 2 weeks before slaughter. The samples of cloacal swabs were collected using swabs (NRS II Transwab swabs with 10 Buffered Peptone Water, Medical Wire & Equipment, Corsham, United Kingdom) in a poultry slaughterhouse. Chickens were raised on 25 unrelated farms located in Greater Poland Voivodship. Birds that were sent to slaughter at 6 weeks of age, weighed an average of 3 kg and belonged to the Ross 308 breed. The chicken meat samples were obtained from neck skin. These samples were delivered to the laboratory for testing, as part Salmonella monitoring program, from 5 different poultry slaughterhouses located in the Greater Poland Voivodeship, from different periods of production.
Information regarding the antibiotics used in the above chickens (name of the antibiotic, withdrawal periods) was included in the food chain documentation. In the treatment of poultry, the most frequently used antibiotics were: Amoxicillin, Enrofloxacin, Doxycycline, Sulfamethoxazole / Trimethoprim.
The samples were placed in buffered peptone water (BioMerieux, Marcy l'Etoile, France) and incubated at 35 °C (± 1 °C) for 18 h (± 2 h) under aerobic conditions. Next, the material was plated on MacConkey agar medium (OXOID, Basingstoke, United Kingdom) and incubated for 24 ± 2 h under aerobic conditions at Table 5. Phenotypes, antibiotic resistance and prevalence of class 1 and 2 integrons, detected in 19 MDR E. coli isolates obtained from meat (Group C).   Table 6. The specificity of the PCR reaction (product length-base pair, bp) was verified by electrophoresis on a 1.5% agarose gel.

Sequencing of the variable regions of Integron 1 and Integron 2.
In the bacterial isolates containing the IntI1 and the IntI2 genes, the variable regions (VRs) of both of the studied integrons were sequenced in order to reveal the specific DNA sequence (plausibly bearing the multidrug resistance genes) within the integron structure. The regions were amplified with a set of specific primers: Integron CL (for integron 1) and Integron CL JJ (for integron 2), sequences are listed in Table 6. The PCR reaction conditions, and primer concentrations were the same as described for IntI1 and IntI2 amplification, with an annealing temperature of 61 °C. The amplified gene cassettes of similar PCR product length (base pair, bp) were sequenced by the Sanger method. Prior to sequencing, the PCR products were purified with FastAP and ExoI enzymes (EF0654 and EN0581, Thermo Fisher Scientific, Waltham, Massachusetts, USA), amplified With the BigDye™ terminator v3.1 Cycle Sequencing Kit (4,337,458, Life Technologies, Carlsbad, California, USA) and purified on Sephadex G50 (G5050, Sigma, St. Louis, Missouri, USA) by filtration. Sequencing of the cassette arrays was done with Applied Biosystems ABI 3130xl 16-capillary array genetic analyzer (Applied Biosystems, Waltham, Massachusetts, USA). Data was analysed with the Seqman software.
Statistical analysis. The significance of the differences between presence and absence of integrons of class 1 and 2, and the predominance of class 1 integrons in relation to class 2 integrons were tested within each group (A, B, and C) using Pearson's Chi-squared Test. Relation between the presence of integrons and the phenotype of multi antibiotic resistance was tested with the use of Poisson Regression analyses where the dependent variable was the number of antibiotics, resistance, and the independent variable was the integron's presence. These analyses were performed separately for integron class 1 and 2 and for each group (A, B, and C). All of the statistical analyses were performed with the use of the R environment 66