Genetic and antimicrobial resistance profiles of non-O157 Shiga toxin-producing Escherichia coli from different sources in Egypt

Background The Shiga toxin-producing Escherichia coli (STEC) represented a great risk to public health. In this study, 60 STEC strains recovered from broiler and duck fecal samples, cow’s milk, cattle beef, human urine, and ear discharge were screened for 12 virulence genes, phenotypic and genotypic antimicrobial resistance, and multiple-locus variable-number tandem-repeat analysis (MLVA). Results The majority of strains harbored Shiga toxin 1 (stx1) and stx1d, stx2 and stx2e, and ehxA genes, while a minority harbored stx2c subtype and eaeA. We identified 10 stx gene combinations; most of strains 31/60 (51.7%) exhibited four copies of stx genes, namely the stx1, stx1d, stx2, and stx2e, and the strains exhibited a high range of multiple antimicrobial resistance indices. The resistance genes blaCTX-M-1 and blaTEM were detected. For the oxytetracycline resistance genes, most of strains contained tetA, tetB, tetE, and tetG while the tetC was present at low frequency. MLVA genotyping resolved 26 unique genotypes; genotype 21 was highly prevalent. The six highly discriminatory loci DI = 0.9138 are suitable for the preliminary genotyping of STEC from animals and humans. Conclusions The STEC isolated from animals are virulent, resistant to antimicrobials, and genetically diverse, thus demands greater attention for the potential risk to human. Supplementary Information The online version contains supplementary material available at 10.1186/s12866-021-02308-w.


Background
Shiga toxin-producing Escherichia coli (STEC) is a pathotype of E. coli that produces Shiga toxins which show molecular similarity to that produced by Shigella dysenteriae type 1 [1]. The STEC serotypes are major foodborne pathogens that cause non-bloody to bloody diarrhea that may ultimately lead to hemolytic uremic syndrome (HUS) [2]. Although ruminants are the main reservoir of STEC that contaminate environment and foods of animal and plant origins, STEC isolates were confirmed in poultry, psittacine pet birds, wild birds and pigeons, dogs, and pigs [3]. Many human outbreaks were caused by the top six STEC serogroups, notably O26, O45, O103, O111, O121, and O145 [4]. The incidence of infections caused by one of the top six STEC serogroups in the USA soared from 0.19 per 100.000 in the year 2007 to 0.79 per 100.000 in the year 2014 [1]. The lipopolysaccharide antigen enables E. coli to colonize bladders and increases bacterial resistance against hydrophobic antibiotics [5,6].
Shiga toxins 1 and 2 are major virulence factors of STEC. There are three subtypes of the stx 1 gene, including stx 1a , stx 1c , and stx 1d and ten subtypes of stx 2 , indicated as stx 2a-2k [7][8][9][10]. Globotriaosylceramide receptors in eukaryotic cell membranes are the binding targets for the Stx1B subunits and for the B subunits of most Stx2 types, whereas; the globotetraosylceramide is the target for Stx2e B subunits [11]. The DNA sequence encoding Stx1 is highly conserved and only a few Stx1 subtypes have been reported; by contrast, Stx2 has numerous subtypes within a range of 84-99% sequence similarity [12,13]. The Stx2 toxin has been associated with severe disease [14]; it is a 1000-times more toxic than Stx1 to renal microvascular endothelial cells, and stx 2c subtype is more commonly reported in patients suffering from HUS [15]. Fecal samples from healthy humans and animals, as well as feces-contaminated environments and foods contain a high burden of STEC strains and phages carrying the stx gene [16]; these can lysogenize nonvirulent bacterial strains and convert them into Shiga toxin-producing types [17]. Intimin is a protein encoded by the bacterial eae gene. It is a component of the enterocyte effacement complex which generates attaching and effacing lesions. The majority of STEC infections that produce serious complications like hemolytic uremic syndrome (HUS) are caused by bacteria that attach to and efface enterocytes (LEE). However, HUS has been linked to a subset of STEC isolates that do not possess LEE [18]. Likewise, enterohemolysin (ehxA) disrupts the cytoplasmic membranes of mammalian cells [14]; intimin and enterohemolysin contribute to STECassociated virulence.
The development and transmission of antimicrobialresistant STEC have become tremendous problems worldwide, as these bacteria may be transferred from poultry to the human food chain [19], and from cattle through farm manure runoff and occupational exposure [20]. Continuous monitoring of antimicrobialresistant STEC provides information on the development of resistant serotypes, the dynamics of bacterial transmission, and antimicrobial abuse [20]. Furthermore, extended-spectrum β-lactamase (ESBL) encoded by the blaCTX-M gene has been identified on plasmids; most hospital-and community-acquired infections with E. coli or Klebsiella pneumoniae are associated with strains that contain CTX-M-type genes [21]. More than 30 different genes encoding for resistance to tetracycline have been identified, involving two major resistance mechanisms: those that promote active efflux of antibiotics and others that prevent interactions with the bacterial ribosomes. The genes tet (A, B, C, D, E, and G) encode control proteins that promote active efflux [22].
Multiple-locus variable-number tandem-repeat (MLVA) analysis has recently emerged as a rapid and highly discriminatory technique for E. coli genotyping [23][24][25]. The methodology detects polymorphisms within variable number tandem repeats (VNTRs) that are present in loci dispersed over the bacterial genome. Some of these VNTRs display critical polymorphisms that can distinguish between highly related clonal strains. MLVA is a promising tool for bacterial genotyping and may even be more effective than pulsed-field gel electrophoresis which is currently the gold standard for E. coli genotyping [23,26]. MLVA was successfully implemented for elucidating the molecular epidemiology of the E. coli O157:H7 strain [27]. Izumiya et al. [28] confirmed the applicability of this method for genotyping of O26 and O111. Researchers can perform efficient STEC subtyping using MLVA, which is a very sensitive tool. Future epidemiological investigations of STEC clonality, including both O157 and non-O157 isolates, could benefit from the diversity present in many serotypes [29].
The purposes of this study were to determine the prevalence of STEC strains among different animal and human samples. And to detect various Shiga toxin genes and their subtypes, phenotypic antimicrobial resistance, extended-spectrum β-lactamases, and tetracycline resistance genes in STEC from diverse sources in a large geographic area in Egypt. To unravel the associations between phenotypic and genotypic features and origin of strains. Moreover, to compare the MLVA profiles with Shiga toxin gene combinations to understand any genetic similarities between animal and human strains. To evaluate the discriminatory power of MLVA loci combinations for elucidating the vital MLVA combination that could be used in preliminary molecular epidemiological studies when rapid results are required.

Sampling, isolation, and identification
During the year 2018, a total of 207 samples were collected, including 15 and 20 fecal swabs from broilers and ducks, respectively, 10 samples of cattle meat, 12 samples of cow's milk, 50 samples of human urine from cases with urinary tract infections, and 100 swabs of human ear discharge from cases with otitis media, the urine samples and ear swabs were collected from separate human cases. For the fecal and meat samples 25 g were collected, while for milk samples 10 ml were collected after discarding the first strips, and 10 ml urine samples were collected from the midstream urine. The human cases are cattle farm workers and farmers rearing broilers and ducks on the small scale. The chicken, duck, and cow's milk samples were collected from farms at Sadat City, Minoufia, Egypt, and these samples were collected after a written informed consent from the owner. This research was performed according to the recommendations of the U.S. Government for the utilization and care of vertebrate animals used in testing, research, and training. The cattle meat samples were collected from 10 butcher shops at Sadat City, and the urine and ear discharge samples were collected from the Central Hospital of Sadat City and Benha University Hospital, respectively. An Informed consent was obtained from all human participants. The adopted methods for handling of human samples were carried out in accordance with relevant guidelines and regulations provided in the Declaration of Helsinki. The fecal, meat, and ear discharge samples were collected in sterile plastic bags and the milk and urine samples were collected in 10 ml sterile tubes. All the samples were labeled, packed, cooled in an icebox, and transported immediately to the Central Bacteriology, Mycology, and Immunology laboratory at the Faculty of Veterinary Medicine, University of Sadat City and stored at 4°C prior to analysis; all samples were processed as fast as possible to optimize the quality of the findings. The isolation and identification procedures of STEC strains comply with the methods outlined by the U. S. Food and Drug Administration Bacteriological Analytical Manual (FDA-BAM) [30]. Briefly, the samples were used to inoculate 225 ml of brain heart infusion broth which was incubated at 35°C for 3 h to promote resuscitation of potentially injured cells. These preenriched samples were then transferred to 225 ml of tryptone phosphate broth and incubated at 37°C for 20 h. One ml of the enriched broth was plated onto Levine's eosin-methylene blue (EMB) and MacConkey (MAC) agar plates which were incubated for 18-24 h at 35°C. Colonies with the characteristic metallic sheen on EMB agar were selected for STEC identification and analysis using biochemical tests. The presumptive STEC colonies (3)(4)(5) were cultured onto citrate utilization (−), triple sugar iron (− H 2 S), and urea agar slants (−ve), and were also subjected to the catalase (+), methyl-red (+), indole (+), and Voges-Proskauer (−) tests; ability of fermenting glucose and lactose sugars (+) was also evaluated. The isolates that exhibited the anticipated results were recorded as STEC isolates [31].
The STEC isolates were subjected to further confirmation using Biolog GN2 and GP2 MicroPlates (Gramnegative and Gram-positive tests, respectively) and the Biolog Microlog 3.70 database and software (Biolog, Hayward, Calif.) [32]. E. coli American type culture collection (ATCC) 11775 and Salmonella enterica subsp. enterica Berta ATCC 8392 were included as positive and negative controls, respectively.
The study design and all the experimental protocols were approved by the Committee for Animal Care and Use, Faculty of Veterinary Medicine, University of Sadat City, Egypt, and the given number was 2018-50. In addition, the committee approved the utilization of human samples within this study after the revision of the informed consent form.

Detection of O-serogroups
The confirmed isolates were re-cultivated on MAC agar and submitted to the Central Laboratories of Ministry of Health, Egypt. Slide agglutination tests were performed using commercial monovalent and polyvalent antisera supplied by Denka-Seiken (Japan).

Antimicrobial susceptibility patterns of STEC strains
The antimicrobial susceptibility patterns of STEC strains were detected after screening the isolates against a panel of 13 antimicrobial agents (Oxoid, UK; Table 1), selected on the basis of their medical importance. One ml of the calibrated bacterial suspension (0.5 McFarland standard units) which include 1.5 × 10 8 colony forming unites/ ml were inoculated on Mueller-Hinton agar; E. coli ATCC 25922 was utilized as the quality control. The Kirby-Bauer disk diffusion method was utilized to determine antibiotic sensitivity; the results were interpreted according to Clinical Laboratory and Standards Institute criteria [33]. The multiple antimicrobial resistance index (MAR) for each strain was determined by dividing the number of antimicrobials to which the strain was resistant by the total number of antimicrobials used.
Prioritization criterion 1 (P1): an antimicrobial used widely among patients with critical infections and in bacterial diseases in health care settings for which this antimicrobial class is the only or one of few alternatives available. Prioritization criterion 2 (P2): an antimicrobial used widely and of the class that may be useful for treating critical infections in health care settings but whose use may favor the generation of resistance. Prioritization criterion 3 (P3): The antimicrobial class typically chosen to control infections in those infected with resistant bacteria or bacteria that harbor resistance genes from non-human origins; NA, no prioritization has been assigned.
The resolution of PCR products of virulence genes, antibiotic resistance genes, and MLVA loci genes were analyzed using the QIAxcel machine at the Department Bacteriology, Mycology, and Immunology, Faculty of Veterinary Medicine, University of Sadat City.

Statistical analysis
The free online calculator for chi-square at https://www. socscistatistics.com/tests /chisquare2/default2.aspx was utilized to evaluate recovery rates of STEC, the frequencies of serotypes, detection of virulence factors, virulence profiles, efficacies of antimicrobials, multiple antibiotic resistance (MAR) indices, and detection of antimicrobial resistance genes. Phenotypic antimicrobial resistance profiles and associated genes confirmed in this study were changed to binary codes for statistical analysis. Sensitivity to given antimicrobial agent recorded as response 0 and resistance was recorded as response 1. The presence or absence of a specific resistance gene was also scored as 1 or 0, respectively. A heatmap, hierarchical clustering, and Pearson correlation coefficient were calculated using the online tools at https://software. broadinstitute.org/morpheus/. The discriminatory index (DI) was calculated according to the formula of Hunter and Gaston [40]; S = Simpson's index of diversity, calculated as.
This calculation assesses the probability that MLVA genotyping will assign two randomly tested unrelated serotypes inside the microbial population to different classifications or groups. The online tools at http://insilico.  ehu.es/mini_tools /discriminatory_power/index.php were used to calculate the DIs of MLVA loci and their combinations. Moreover, the 95% confidence interval for each DI result was calculated using the free online tools at: http://www.compa ringp artit ions.info/?link=ToolI . A dendrogram based on the MLVA-associated diversity was constructed using the BioNumerics software v. 6.6 (Applied Maths, Sint-Martens-Latem, Belgium).
In this first set of tested antimicrobials the resistance to nalidixic acid and chloramphenicol was more prevalent in the isolates, compared to other antibiotics. While resistance to cephradine, erythromycin, norocillin, oxytetracycline, clindamycin, streptomycin, and penicillin G represented 65, 73.3, 75, 80, 81.7, 91.7, and 96.7%, respectively, for this group of antimicrobials there noticed increased resistance compared with the aforementioned types. We detected significant differences among the efficacies of these antimicrobials with p < 0.05 (Supplementary Table 2 and Supplementary Fig. 2). For broiler and duck strains, the MAR index ranges were 0.38-0.77 and 0.23-0.92, respectively. Furthermore, the MAR index for cattle strains range was 0.38-0.92 and from human urine and ear discharge, the ranges were 0.15-0.85 and 0.31-0.92, respectively. The strains O78 (sample #17), O91:H21 (sample #30), and O17:H18 (sample #58) from duck, cattle, and human ear discharge exhibited the highest MAR indices at 0.92 for each, with a significant difference p < 0.05 was present among the dissimilar indices of different strains (Supplementary Table 3).

Associations between isolation source, strain, and phenotypic and genotypic characters
The relationship of the existence of antimicrobial resistance, resistance genes, strain, and the source of samples was discovered to detect possible associations among the isolates. Phenotypic antimicrobial resistance profiles and associated genes confirmed in this study were changed to binary codes for statistical analysis. Sensitivity to given antimicrobial agent recorded as response 0 and resistance was recorded as response 1. The presence or absence of a specific resistance gene was also scored as 1 or 0, respectively. The Pearson correlation coefficient was calculated using the online tools at https://software. broadinstitute.org/morpheus/. Some strains from broilers and ducks exhibited high virulence, phenotypic and genotypic antimicrobial resistance (Fig. 1). There was a significant difference (p < 0.05) in strains impact on virulence genes, phenotypic antimicrobial resistance, and resistance genes (Fig. 2). Correlation matrix analysis ( Fig. 1) and hierarchical clustering with heat map (Fig. 2) were utilized to detect associations between the phenotypic and genotypic features and origin of the strains. Correlation analysis showed positive relationships among the presence of resistance to β-lactams especially to ampicillin and the presence of β-lactamase genes as the blaCTX-M-1 and blaTEM (Fig. 2, p < 0.05). Significant positive correlations of antibiotic resistances proved co-occurrence of resistance may be predominant, (p < 0.05) and confirmed the presence of multiple-drugresistant strains (MDR). For example, resistance to amoxicillin/clavulanic acid and ampicillin were positively correlated with resistance to amikacin, cephradine, clindamycin, doxycycline, and nalidixic acid resistances tested. The existence of the blaCTX-M-1 and blaTEM were positively correlated with resistance to chloramphenicol and nalidixic acid (p < 0.05).

Analysis of the MLVA loci
The data presented in Supplementary Results  < 0.3). The selected loci were highly predictive for discrimination among strains with a 95% confidence interval (CI) 1.000-1.000.

Evaluation of the discriminatory power of MLVA loci combinations
The 60 STEC strains were resolved to 26 different genotypes with a DI of 0.9277 using 8 MLVA loci, which were clustered into 24 diverse clustered types, five unique types, and 19 clustered types. The clusters contained 1 (n = 1) to 2 (n = 12) analogous strains (Supplementary Table 8), and the 23 diverse types were classified into two groups (Fig. 3). Group 1 was less complex and included nine clustered types; by contrast group 2 was more complex and contained fifteen clustered types. The discrimination power was compared among different MLVA combinations. The first combination included VNTR 3,9,17,19,36

Discussion
The STEC strains such as O157 and non-O157 have acquired genetic traits that made them potential risk for human. In humans, STEC causes a wide range of infections including diarrhea, complicated hemorrhagic colitis Fig. 2 Heatmap and hierarchical clustering of E. coli isolates to 5 clusters based on their phenotypic (antimicrobial resistance), genotypic (antimicrobial resistance genes), and virulence genes expressing differences between isolates. Red represented presence and blue represented absence of phenotypic resistance, resistance genes, and virulence genes. Hierarchical clustering was performed using Wald's method and a binary distance matrix (HC), HUS, and thrombotic thrombocytopenic purpura. Humans may acquire infection as a result of contamination of food and water [41,42]. Our first goal was to identify STEC from multiple sources. Our most prominent source was duck fecal samples; STEC strains were isolated from 14/20 (70%) samples. These results are nearly similar to published results from Malaysia during 2012 [43], which confirmed the existence of STEC in duck feces with a frequency of 87.93%. Moreover, we isolated STEC from 50% of the broilers; this result was smaller than reported results from Egypt during 2015 [44], which proved that E. coli was present in 75% of local broilers. Our isolation rate from cow's milk was 50% similar to results from Egypt during 2019 [45], which confirmed the existence of E. coli in 50% of the composite milk samples from both cows and buffaloes. We isolated STEC from 50% of the cattle meat samples, which was higher than Gwida et al. [46] who identified E. coli in 27% of raw beef. These results confirmed that the STEC could be isolated from a wide variety of food animals and poultry found in small farms within rural localities [47]. Likewise, our STEC isolation from human urine samples was lesser than Córdoba et al. [48], who identified E. coli from patients of suspected urinary tract infections in primary care, at Denmark 2017. The isolation from human ear discharge was lesser than Kibret and Abera [49]. Among the strains from broiler and duck, O78 was the most prevalent, detected in 6/60 (10%) which was similar to Wang et al. [50] and Abd El Tawab et al. [51] who confirmed the high prevalence of O78 among strains that cause avian colibacillosis. Both O2 and O128 strains were previously reported in poultry by Byomi et al. [52] and in ducks as reported by El-Shabrawy et al. [53], moreover the existence of O121:H7 and O146 strains in chicken was confirmed by Enany et al. [54], while recorded in ducks by Wang et al. [50]. The distribution of O1, O26:H11, O91, O127, and O153 strains in samples from chickens and ducks were similar to Abd El-Mongy et al. [44] and Wang et al. [50]. The O26:H11, O91:H21, O121:H7, O128:H2, O153:H2, O1: H7, O2:H6, O78, O127:H6, and O146:H21 [50][51][52][53][54][55] are serious avian STEC strains that cause severe localized or extraintestinal infections. The extraintestinal infection called colibacillosis which affects broiler chickens from 4 to 6 weeks of age and distinguished by causing acute fatal septicemia or sub-acute airsacculitis, fibrinous pericarditis, peritonitis, and salpingitis [54]. The O26:H11, O111:H2, and O128:H2 strains have all been identified previously in cattle meat products [56]. These strains from meat are implicated in the etiopathogenesis of calve diarrhea either mild or sever with significant alterations in the animal hematological and biochemical parameters, moreover these strains exhibited public health importance [57]. Furthermore, the isolation of O55:H7, O86, O91:H21, and O127:H6 from cases of bovine mastitis worldwide was confirmed [58]. Interestingly, the isolation of O2:H6, O15:H2, O7:H2, O17:H18, O8:H21, O83, O125:H21, O75, and O124 in association with human urinary tract infections suggests that these strains may be uropathogenic [59]. We found no published reports documenting the serotyping of STEC associated with otitis media; here, we found that O8:H21, O15:H2, and O17:H18 were the most prevalent serotypes. The non-O157 STEC are more frequently isolated with about 4-fold higher isolation rate than O157, from HUS cases the non-O157 STEC proportion varying from 7 to 90% [60]. The enteropathogenic E. coli serogroups are human-specific pathogens which cause endothelial infection, intestinal thrombotic microangiopathy, and ischemic enteritis [61]. The O1:H7, O2:H6, O7:H2, O8:H21, O15:H2, O17:H18, O26:H11, O75, O83, O91:H21, O121: H7, O124, O125, O153:H2, and O128:H2 strains imparted zoonotic impact and have been isolated from human infections [62]. The O55, O86, O111, and O127 strains have been considered as major causes of acute and persistent infantile diarrhea in many developing Fig. 3 Dendrogram of STEC strains based on their multiple-locus variable-number tandem-repeat analysis (MLVA) profiles. Strains are clustered into two groups, group 1 less complex and included ten types; and group 2 more complex and contained sixteen clustered genotypes countries [62,63]. The O1:H7, O2:H6, O8:H21, O15:H2,  O55:H7, O75, O91:H21, O111:H2, and O128:H2 STEC strains have been isolated from HUS patients [62]. The Shiga toxin gene, stx 1 , was detected in all strains, while stx 2 was detected in 80% of strains; the increased detection rate from broilers, ducks, beef, and human urine samples agree with the recent report from Egypt during the year 2020 [64]. Moreover, the high existence patterns of both stx 1 and stx 2 from cattle milk strains were confirmed by Ranjbar et al. [65]. Of the stx 1 subtypes, stx 1d was detected most prominently in 85% of the isolates, while stx 1c subtype was not detected at all. Of all the stx 1 positive strains, stx 1d subtype was detected predominantly in our isolates from chicken and duck (100%) and was identified in 73.7 and 72.7% of the isolates from humans and cattle, respectively. From these results, the stx 1d was commonly found in animal strains as confirmed by Kumar et al. [66], while its high distribution among human strains wasn't confirmed by him, but he stated that it could be linked with a mild course of disease. Moreover, the high distribution of the stx 1d subtype in human strains come in contradiction with that reported by EFSA BIOHAZ Panel et al. [67], those confirmed that the stx 1d subtype wasn't predominantly associated with hospitalisations and bloody diarrhea. As the human cases were cattle farm workers and farmers rearing broilers and ducks on the small scale, this result could be regarded to the mobile genetic elements like bacteriophages, insertion sequence elements, pathogenicity islands, plasmids, and transposons that play a vital role in the evolution of human STEC and changing them to stx 1d subtype positive [67]. The stx 2 subtypes were screened and stx 2c subtype prevalence was 7.  [70], which confirmed that chicken E. coli isolates harbored no stx 2e subtype. Furthermore, the distribution of this subtype in cattle meat and milk STEC exceeded that recorded in a report from Germany during the year 2011 [71], while its existence in human cases was confirmed by EFSA BIOHAZ Panel et al. [67]. The eaeA was detected in 3.3% of isolates, which was smaller than Elsayed and Munir [64]. Although we used a universal oligonucleotide primer pair EAE-F and EAE-RB with homology to the 3'variable region of eae (that detects all types of eae variants described at the moment of the manuscript published by Blanco et al. [36], who screened 514 STEC isolates for stx and eae genes and their subtypes, there found that most of our STEC isolates harbored no eae gene which shows similarity to his results. But two O26:H-of his STEC isolates harbored eae type β1, while one isolate O26:H11 contained no eae that confirms the possibility of existence of O26:H11 negative for eae, added to that, he confirmed three isolates of O111:H-contained γ2 type of eae, which is quite different from from our O111:H2. It is clear that most STEC infections with serious complications like hemolytic uremic syndrome are caused by bacteria that attach to and efface enterocytes and harbor active eae gene. However, HUS has been connected to a subset of STEC isolates that do not possess eae genes [18]. The ehxA was detected in 77% of isolates, a rate surpassed that of Elsayed and Munir [64]. The existence of Shiga toxin genes and virulence genes in most of isolated STEC strains from animals and human represent a serious problem to public health as these strains could result in dangerous infections and could transmit these characters to other pathogenic and nonpathogenic bacterial agents through the mobile genetic elements [17,67]. Of note, to the best of our knowledge, this is the first report of the specific prevalence of given combinations of the stx genes; then 10 combinations featured here are novel and not previously reported [14,66], these combinations were recorded to express the more frequently found stx genes among the gained STEC strains which can differentiate between them. From the distribution patterns of stx gene combinations, it was clear that several strains harbored many copies of stx genes, as some combinations composed of 4 copies as stx 1 , stx 1d , stx 2 , stx 2e which represented 51.7% of the strains and 5 as stx 1 , stx 1d , stx 2 , stx 2c , stx 2e which represented 1.7% of the obtained strains. Although our stx gene copies were high, this fact comes in agreement with the published data from Finland during 2002 [15], which confirmed the existence of 11 stx gene combinations; the most prevalent combinations were stx 2 with stx 2c estimated (42%) which contained two copies. And from the O157 strains, 64% carried stx 2 with stx 2c versus 2% of the STEC strains. Furthermore, the existence of multiple copies of stx genes in human strains was confirmed by a report from Germany during 2006 [72], which proved that the genotype stx 2dactivatable , stx 1 , and stx 2 that contains 3 copies of stx genes was present in 8/60 (13.3%) of STEC isolates. The STEC strains were screened against a panel of 13 antimicrobial agents that were selected on the basis of their medical importance. Prioritization criterion 1 (P1): an antimicrobial used widely among patients with critical infections and in bacterial diseases in health care settings for which this antimicrobial class is the only or one of few alternatives available. Prioritization criterion 2 (P2): an antimicrobial used widely and of the class that may be useful for treating critical infections in health care settings but whose use may favor the generation of resistance. Prioritization criterion 3 (P3): The antimicrobial class typically chosen to control infections in those infected with resistant bacteria or bacteria that harbor resistance genes from non-human origins [73].
Our isolates showed varying susceptibilities to antimicrobial agents. Among our findings, 73% of the isolates were resistant to erythromycin which was lower than Rubab and Oh [74]; by contrast, the rates of resistance to nalidixic acid, oxytetracycline, and streptomycin were higher than him. The extent of resistance to ampicillin was lesser than Rubab and Oh [74], and the rate of resistance to chloramphenicol outpaced that of Elsayed and Munir [64]. The resistance observed to cephradine, norocillin, and penicillin G exceeded that of Elsayed and Munir [64], and the resistance to doxycycline was lower than him. The extent of resistance to amoxicillin/clavulanic acid was smaller than Elsayed et al. [38] and the resistance to clindamycin was greater than Rubab and Oh [74]. Several isolates expressed multidrug-resistant (MDR) and extensively drug-resistant (XDR) phenotypes. The correlation analyses proved co-occurrence of resistance to various antimicrobials, exemplifies a significant concern for animal and human medicine alike. Moreover, resistance to some antimicrobials was linked with susceptibility to others. As clear, erythromycin resistance was related with susceptibility to amikacin, amoxicillin/clavulanic acid, chloramphenicol, doxycycline, and oxytetracycline. This finding is remarkable because when discussing MDR or XDR STEC it may promote the selection of alternative antimicrobials. Most of the strains exhibited MAR indices that surpassed 0.2; these results suggest that these serotypes may have originated from high-risk sources with uncontrolled implementation of antimicrobial agents [75]. Inadquate antimicrobial selection and abuse can lead to resistance in different bacteria and make it more difficult to treat bacterial infections [76]. High frequencies of antimicrobial drug resistance were observed in STEC strains recovered from the collected samples, which was confirmed to be more common in non-O157 isolates and could contribute to serious disease outcomes [77]. Many antimicrobials were implemented for food producing animals as growth promoters and for prevention, control, and treatment of diseases. From these types, tetracyclines, penicillin, and cephalosporins were used, that represent public health hazards. This uncontrolled use of antimicrobials results in emergence of antimicrobial resistance, hypersensitivity, carcinogenicity, bone marrow depression, mutagenicity, teratogenicity, and disturbance of intestinal normal flora [78,79]. The extended-spectrum β-lactamases (ESBLs) genes mediate the production of enzymes that destroy antimicrobials belonging to the penicillin and cephalosporin classes and turn them ineffective [80]. The bacterial pathogens can carry antimicrobial resistance genes on mobile elements. These elements can be horizontally transferred to various bacterial species, which could change the recipient strain to become drug resistant. The Kluyvera spp. are environmental saprophytes that considered the source of bla CTX-M . The CTX-M-type genes are the most common ESBL genes found in E. coli and Klebsiella pneumoniae isolates responsible for the worldwide community-and hospital-acquired infections. Moreover, the TEM-genes are also found in various types of environmental saprophytes [21].
The blaCTX-M-1 was detected in 13.3% of isolates, which was confirmed in 2/7 (28.6%) of broilers and was lower than the published results by Kim et al. [81]. It was detected in 2/14 (14.3%) of duck strains which was lesser than recent records from India during 2020 [82], while its existence in cattle meat comes in contradiction with Kennedy et al. [83], who proved its absence in STEC strains from abattoir. Additionally, its presence in the STEC from raw milk was confirmed by Ahmed et al. [84]. The ampicillin-resistance gene blaTEM was found in 2/7 (28.6%) of the stx positive isolates of broilers that contradicts with Saad et al. [85], who confirmed its existence in stx negative E. coli from chicken, while the existence of blaTEM in STEC strains from duck was confirmed by Kim et al. [81].
The selected tetA, B, C, D, E, and G oxytetracycline resistance genes encode for efflux proteins present in the cytoplasmic membrane of gram-negative bacteria. They work as antiporters exchanging a proton for a monocationic magnesium-tetracycline complex and reducing the tetracycline amount in the bacterial cytoplasm [86,87]. In Egypt, there found low number of reports that examined the existence of tetracycline resistance genes in E. coli isolated from a large variety of animals and human samples with various histories of exposure to tetracyclines. In our study, the tetA and B exhibited high distribution patterns among animal and human strains as confirmed by Bryan et al. [88]. The existence of tet C in chicken comes similar to Bryan et al. [88], while its existence in duck isolates differs. Although we gained high distribution patterns of tet E and G among animal and human strains, tetD was absent, these results come in contradiction with Bryan et al. [88], who found no tet E and G among animal and human isolates and confirmed existence of tetD gene among human isolates. Although Bryan et al. [88] found that 22.2 and 1.9% of the isolates contained two and three tet genes, respectively, our results surpassed him in the number of contained genes as we found that 7/60 (11.7%) harbored two genes, 20/60 (33.3%) expressed three genes, 18/60 (30%) harbored four genes, and 2/60 (3.3%) contained five genes. And the rational interpretation of this fact could be regarded to the strong selection pressures of environments contained high levels of tetracycline that results in acquisition of several tetracycline resistance genes [88]. In this study the identification of resistance genes emphasizes the fact that STEC can serve as reservoirs for antimicrobial resistance genes that could be passed to pathogenic microorganisms that infect humans.
This study is one of the first to consider the applicability of using the eight MLVA loci described by Izumiya et al. [28] for molecular characterization and genotyping of STEC strains in Egypt. The data presented add to our understanding of the genetic diversity and relatedness of various STEC strains. The methodology will facilitate comparisons among distinct genetic profiles with respect to the origin of host and geographic locations. The MLVA analysis technique was chosen based on technical and financial considerations as well as the high discriminatory power [23,26]. The VNTR 3,9,17,19, and 36 loci were the most polymorphic and permitted us to evaluate our strains with increased resolution. VNTR 3,9,17,25, and 36 exhibited high discriminatory indices and allele numbers when compared with the same loci evaluated in E. coli strains O26, O111, and O157 by Izumiya et al. [28]. The DI of VNTR 34 was similar to that of Izumiya et al. [28] although the allele number was lower. Here, VNTR 19 exhibited a greater DI and a lesser allele number, VNTR 37 exhibited a smaller DI and an increased allele number compared with the same loci evaluated by Izumiya et al. [28]. The MLVA typing using the eight aforementioned loci VNTR 3,9,25,17,19,34,36, and 37 provided discriminatory genotyping for the 60 strains. Our results revealed 26 distinct genotypes with different allelic profiles; the efficacy of this methodology surpassed that based on stx gene combinations. As such, our results concur with those of previous studies as we found that MLVA is capable of high discriminatory power that surpasses the serotype grouping via PCR amplification of virulence genes [89].

Conclusions
In conclusion, the STEC strains are widely distributed in broilers, duck, cattle, and human infections. This is the first report of a detailed detection of virulence repertoire, phenotypic and genotypic antimicrobial resistance, and MLVA typing of STEC strains from different sources in Egypt. The gained isolates showed lowered resistance to amikacin, amoxicillin/clavulanic acid, doxycycline, ampicillin, nalidixic acid, and chloramphenicol. Furthermore, the uncontrolled use of antibiotics for STEC infections in animals represents a potential risk for public health. There was a significant impact of serogroups and serotypes on virulence genes, antimicrobial resistance, and resistance genes. MLVA typing considered useful genotyping method and the results from the VNTR 3,9,17,19,36, and 37 loci are sufficiently robust that they can be used for preliminary molecular epidemiological studies. Most of the animal STEC strains were not found in human infections except the O2:H6 which expressed different MLVA profiles. The obtained results will be useful toward controlling STEC; with these methods it will be comparatively easy to assess linked clusters, cluster growth, and transmission dynamics.
Additional file 1 S1. Results of distribution patterns of virulence genes among various E. coli strains. S2. Antimicrobial susceptibility patterns of 60 E. coli strains from different sources. S3. Results of multiple antimicrobial resistance indices of various strains from different sources. S4. Results of distribution of class 1 and 2 integrons, extended-spectrum β-lactamase, and ampicillin-resistance genes in various strains. S5. Results of the distribution patterns of oxytetracycline resistance genes among the phenotypically resistant strains. S6. Results of the gained tandem repeats with the utilized MLVA loci and the discriminatory index. S7. Results of genotypes and allelic profiles of various STEC strains after MLVA typing. S8. Evaluation of the discriminatory power of different MLVA loci combinations.

Funding
This research received no external funding.

Availability of data and materials
The data and material are available in the manuscript and the supplementary material.

Declarations
Ethics approval and consent to participate The study design and all the experimental protocols were approved by the Committee for Animal Care and Use, Faculty of Veterinary Medicine, University of Sadat City, Egypt, and the given number was 2018-50. In addition, the committee approved the utilization of human samples within this study after the revision of the informed consent form. An Informed consent was obtained from all human participants. The adopted methods for handling of human samples were carried out in accordance with relevant guidelines and regulations provided in the Declaration of Helsinki. The