Phenotypic and molecular characterizations of multidrug-resistant diarrheagenic E. coli of calf origin

Escherichia coli has become one of the most important causes of calf diarrhea. The aim of this study is to determine the patterns of antimicrobial resistance of E. coli isolates from six cattle farms and to identify prominent resistance genes and virulence genes among the strains isolated from the diarrhea of calves. Antimicrobial susceptibility tests were performed using the disk diffusion method, and PCR was used to detect resistance and virulence genes. The prevalence of multidrug resistant (MDR) E. coli was 77.8% in dairy cattle and 63.6% in beef cattle. There were high resistance rates to penicillin (100%, 100%) and ampicillin (96.3%, 86.4%) in E. coli from dairy cattle and beef cattle. Interestingly, resistance rate to antimicrobials and distribution of resistance genes in E. coli isolated from dairy cattle were higher than those in beef cattle. Further analysis showed that the most prevalent resistance genes were blaTEM and aadA1 in dairy cattle and beef cattle, respectively. Moreover, seven diarrheagenic virulence genes (irp2, fyuA, Stx1, eaeA, F41, K99 and STa) were present in the isolates from dairy cattle, with a prevalence ranging from 3.7% to 22.22%. Six diarrheagenic virulence genes (irp2, fyuA, Stx1, eaeA, hylA and F41) were identified in the isolates from beef cattle, with a prevalence ranging from 2.27% to 63.64%. Our results provide important evidence for better exploring their interaction mechanism. Further studies are also needed to understand the origin and transmission route of E. coli in cattle to reduce its prevalence.


Introduction
Diarrheagenic E. coli (DEC) is a significant cause of gastroenteritis and a major health problem in animals and humans. E. coli infection in calves usually causes a variety of clinical signs, including diarrhea, respiratory infections, and sepsis, and then death due to dehydration and exhaustion because of the difficulties in treatment. Previous studies have shown that diarrhea is the most common problem in young calves, causing more than 52% of deaths in unweaned calves (Diarra et al. 2009). In cattle farms, antimicrobials are the most important therapy for bacterial infection. In dairy cattle farms worldwide, periodic treatment of mastitis after bacterial infection is very common, which not only easily leads to bacterial resistance but also raises concerns about the emergence of multidrug resistant (MDR) bacteria (Yang et al. 2021). The use of antimicrobials to treat infections in beef cattle can increase prevalence of antimicrobial resistance (AMR) in enteric pathogens (Cazer et al. 2017). In addition, antimicrobials are frequently used as growth promoters and preventive agents, which further increases the risk of E. coli resistance (Sivaraman et al. 2020). AMR in bacteria of animal origin is considered a major challenge to veterinary medicine and public health (Anes et al. 2020), which not only seriously affects the healthy development of cattle breeding industry but also poses a serious threat to food safety. E. coli has also been used as a sentinel organism for monitoring AMR (de Moyaert et al. 2014). Hence, monitoring AMR in cattle is important to human and animal health.
Some pathogenic E. coli strains use different virulence factors to colonize the hosts' small intestine, avoiding immune response and stimulating the deleterious inflammatory response to produce diarrhea (Croxen and Brett Finlay 2010). Virulence genes that play significant roles in E. coli pathogenicity are associated with diarrhea in animals and humans have been described (Fröhlicher et al. 2008;Huehn et al. 2010). Among the many virulence genes identified in E. coli isolates from cattle, Shiga toxins (Stx1 and Stx2), Yersinia high pathogenicity island (irp2 and fyuA) and intimin (eaeA) were the most significant genes with great public health concerns (Momtaz et al. 2012;Olsson et al. 2003;Momtaz et al. 2013a, b). Cattle are a major reservoir of E. coli, particularly Shiga toxin-producing E. coli (STEC) O157:H7. In addition, E. coli has many serotypes, among which E. coli O157 can cause hemorrhagic colitis and hemolytic uremic syndrome (Iweriebor et al. 2015). Heat-labile enterotoxins (LT) and heat-stable enterotoxins (STa or STb) are the two most important virulence factors responsible for severe diarrhea in cattle (Nguyen et al. 2011;Kumar et al. 2013). The most important adhesins involved in E. coli host colonization are fimbriae. Wellcharacterized fimbriae of E. coli isolated from animals include F4 (K88), F5 (K99), F6 (987P), F41 and F18, are associated with E. coli pathotypes (Maciel et al. 2019). Previous studies have also shown that the ability of E. coli to acquire many different virulence factors may lead to the emergence of invasive strains, which pose a threat to human and animal health (Mellmann et al. 2011). Therefore, the aim of this study is to characterize AMR and identify different resistance genes and virulence genes in E. coli strains isolated from dairy cattle and beef cattle to provide a reference for clinical practice.

Results
Prevalence of AMR in E. coli isolated from dairy and beef cattle A total of 71 E. coli isolates were obtained, including 27 isolates from dairy cattle and 44 isolates from beef cattle diarrheal fecal samples. Subsequently, susceptibility to 15 different antimicrobials was determined for these 71 E. coli isolates. All 27 E. coli isolates from dairy cattle were resistant to penicillin, followed by ampicillin (96.3%), amoxicillin and sulfamethoxydiazine (81.5%), tetracycline and compound sulfamethoxazole (77.8%), with the lowest resistance rate being observed for florfenicol (33.3%) (Fig. 1). Meanwhile, all isolates were sensitive to polymyxin B (100%). Consistent with the results of dairy cattle, the most sensitive antimicrobial was also polymyxin B in the 44 isolates from beef cattle (Fig. 2). The highest resistance rate was also observed for penicillin (100%), which may be related to the widespread use of penicillin for the treatment of E. coli disease. Further analysis showed that the resistance rate of E. coli to antimicrobials (except for florfenicol and polymyxin B) in dairy cattle was higher than that in beef cattle.

Prevalence of multidrug resistant (MDR) E. coli
Multidrug resistance was defined as resistance by an isolate to at least three antimicrobials of the panel belonging to different classes. Resistance of E. coli to seven different types of antimicrobials were analyzed. The results showed that multidrug resistance rates were 77.8% (21/27) in dairy cattle and 63.6% (28/44) in beef cattle. Most isolates from dairy cattle and beef cattle were resistant to five or six different types of antimicrobials. The prevalence of resistance to five different types of antimicrobials was 37% (10/27) in dairy cattle and 18.2% (8/44) in beef cattle. Compared with the isolates from dairy cattle, isolates from beef cattle had a higher prevalence of resistance to six different types of antimicrobials [dairy cattle 29.6% (8/27) vs. beef cattle 31.8% (14/44)] (Table 1). One isolate from beef cattle was resistant to all antimicrobials.

Prevalence of resistance genes in E. coli
Prevalence of 12 different resistance genes was analyzed in E. coli isolates from dairy cattle and beef cattle origins. The results showed that seven different resistance genes were detected in over 50% isolates from dairy cattle (Table 2). Resistance genes that had the highest positive rate were bla TEM (100%), followed by floR, tet (A), , aac (3′)-IIa and sul2. Resistance gene with the lowest positive rate was aadB (0%). However, detection rate of seven drug resistance genes in 44 isolates from beef cattle was over 56%, with 100% positive rate of aadA1, followed by bla TEM , tet (A), and tet (B) ( Table 2). Overall, the positive rates for bla TEM , aadA1, tet (A), tet (B), floR and sul2 were relatively high in the E. coli isolates of both dairy and beef cattle. Consistent with the AMR results, detection rate of resistance genes in dairy cattle was higher than that in beef cattle.
Correlation between the resistance phenotype and resistance genes Consistency analysis of resistance phenotypes and resistance genes to 11 antibiotics showed that β-lactam (penicillin) resistance phenotype had the highest consistency with β-lactam resistance genes (beef cattle K = 1), followed by compound sulfamethoxazole (beef cattle K = 0.59), gentamicin (beef cattle K = 0.56) and florfenico (beef cattle K = 0.41). In dairy and beef cattle, tetracycline resistance phenotype had the lowest consistency (K = − 0.55, K = − 0.77) with tetracycline resistance gene tet (C). Some isolates presenting drug resistance carried resistance genes, whereas some isolates carried resistance genes without manifesting a resistance phenotype ( Table 3).

Coexistence of virulence and AMR genes in E. coli
Further study showed that 49 E. coli isolates carried at least one virulence gene, including 38 isolates from beef cattle and 11 isolates from dairy cattle. Subsequently, the coexistence of virulence genes and AMR genes in these 49 E. coli isolates were analyzed. The results showed that there were at least 4 AMR genes in the isolates containing virulence genes and up to 10 AMR genes (Table 5) in other isolates. Interestingly, all 49 E. coli isolates contained bla TEM and tet (A) genes. In addition, most of 38 isolates from beef cattle contained bla TEM , tet (A), tet (B) and floR   genes, while 11 strains of isolates from dairy cattle carried aac(3′)-IIa and sul2 (Table 6).
Frequency of virulence gene occurrence in isolated E. coli strains exhibiting antimicrobial resistance The frequencies of virulence gene occurrence in isolated E. coli strains exhibiting antimicrobial resistance were detailed in Table 7. The majority of β-lactam-, aminoglycoside-, tetracycline-, sulfonamide-, fluoroquinoloneand chloramphenicol-resistant beef cattle E. coli isolates (more than 50%) were positive for irp2 and fyuA genes with a significant association. Significant associations between the rest of virulence genes and antibiotic resistance were not observed.

Discussion
The emergence and spread of AMR bacteria have become a growing problem and a threat to global public health (WHO 2017). In veterinary practice, penicillin, ampicillin, florfenicol, sulfadiazine, streptomycin, gentamicin and tetracycline are all commonly used antimicrobials for treating E. coli-associated infections. Previous studies showed that all 100 E. coli isolates from Irish cattle farms were resistant to streptomycin, with a resistance rate of 100%, followed by resistance rates of 99% for tetracycline, 98% for sulfonamides, and 82% for ampicillin (Karczmarczyk et al. 2011 (Barigye et al. 2012). In contrast, we found that E. coli isolated from beef and dairy cattle were both susceptible to polymyxin B. These results indicated that E. coli with different origins may have undergone different evolutionary processes and thereby acquired different resistance genes. Interestingly, this research showed that the resistance rate of E. coli to antimicrobials (except for florfenicol and polymyxin B) from dairy cattle was higher than that of beef cattle. Multidrug resistance analysis showed that most isolates from dairy cattle and beef cattle were resistant to five or six types of antimicrobials. Similarly, multidrug resistance rate in E. coli isolated from dairy cattle is higher than that isolated from beef cattle. In dairy cattle, periodic treatment of mastitis after bacterial infection is very common, and antimicrobials are the most important therapies for bovine mastitis, which may be one potential reason for the high resistance rate of E. coli from dairy cattle (Call et al. 2008;Mazurek et al. 2013). Meanwhile, these results suggested that more rational use of antimicrobials in cattle farms was needed to prevent the development of AMR in E. coli. E. coli resistance genes bla TEM and bla SHV were the first described extended spectrum β-lactamase (ESBL) genes in the 1980s, and they were predominant until 2000 (Poirel et al. 2018). Currently, the production of ESBL, especially bla TEM , is one of the most important mechanisms of AMR from the clinical and epidemiological point of view (Poirel et al. 2018). Indeed, previous studies reported that bla TEM was detected in 78.94% isolates from dairy cattle farms in the Nile Delta in Egypt, whereas bla SHV and bla OXA were detected only in 0.87% isolates (Braun et al. 2016). In China, previous studies  (Yang et al. 2018). In the present study, 27 E. coli isolates from dairy cattle farms were tested and it was found that detection rate of bla TEM was as high as 100%, and detection rates of bla SHV and bla OXA were also higher than previous studies. Similar to the results in dairy cattle, 44 E. coli isolates from beef cattle also showed the highest detection of bla TEM (97.7%). In addition, a previous study reported the resistance rates of bla SHV (0%) and bla OXA (0%) in Japanese beef cattle (Yamamoto et al. 2014), while they were 9.1% and 6.8% in this work, respectively. These results indicated that bla TEM was still the most common AMR gene in China and other countries regardless of whether the isolates were from dairy or beef cattle. Furthermore, it is worth noting that detection rates of bla SHV and bla OXA may have a tendency to increase. This research further showed that chloramphenicol and aminoglycoside resistance genes were present in E. coli isolates. Detection rates of floR in dairy cattle and beef cattle were 96.3% and 59.1%, respectively, which were similar to previous reports (Belaynehe et al. 2018;Wu et al. 2011). In addition, aminoglycoside (resistance) genes aadA1 and aadB were detected in 70.4% and 0% of 27 E. coli isolates from dairy cattle and in 100% and 9.1% of 44 E. coli isolates from beef cattle. In Ireland, aadA1 and aadB were identified in 19% and 1% of 100 (MDR) E. coli isolates recovered from dairy cattle (Karczmarczyk et al. 2011). In Iran, aadA1 was detected in 26.2% of E. coli isolates from dairy cattle (Jamali et al. 2018). In Mexico, aadA1 was detected in 17% of E. coli isolates from beef cattle (Martínez-Vázquez et al. 2018). Detection rate of aadA1 in this study is much higher than that reported in other countries. Interestingly, the detection rate of aac(3′)-IIa that has not been reported in previous studies was 56.8% in beef cattle and 96.3% in dairy cattle, which is worth further investigation detection rate of tetracycline resistance gene tet (A) was 97.7%, followed by tet (B) (95.5%) and tet (C) (0%) in 44 E. coli isolates from beef cattle. In isolates from dairy cattle, detection rate of tet (A) was 96.33%, followed by tet (B) (70.4%) and tet (C) (7.4%). sul1 gene was detected in 40.7% and 47.7% while sul2 gene was detected in 96.3% and 77.3% of E. coli isolates from dairy cattle and beef cattle, respectively. These results are similar to those previously reported data (Karczmarczyk et al. 2011;Belaynehe et al. 2018;Shin et al. 2015;Navajas-Benito et al. 2017). Further analysis found that the overall detection rate of resistance genes in dairy cattle was higher than that of beef cattle, suggesting the widespread resistance of E. coli in dairy cattle. Totally 14 different virulence genes were analyzed in E. coli isolates from dairy cattle and beef cattle. However, only 7 diarrheagenesis-associated virulence genes (irp2, fyuA, stx1, eaeA, F41, K99 and STa) were detected in isolates from dairy cattle, and 6 diarrheagenesisassociated virulence genes (irp2, fyuA, Stx1, eaeA, hylA and F41) were detected in isolates from beef cattle. In beef cattle, 28 out of 44 E. coli isolates were positive for irp2 (63.64%), and 27 were positive for fyuA (61.36%). Detection rates of irp2 and fyuA in isolates from dairy cattle were also the highest, both at 22.22%. These results suggested that irp2 and fyuA in E. coli isolates from dairy cattle and beef cattle were the main virulence genes, which was similar to the results of previous studies (Ewers et al. 2004;de Verdier et al. 2012). The results also indicated that detection rate of the main virulence genes irp2 and fyuA in isolates from beef cattle was higher than that in isolates from dairy cattle. Furthermore, detection rates of F41 and eaeA genes were not significantly different between beef and dairy cattle, which was consistent with the results of previous peports (Andrade et al. 2012;Hornitzky et al. 2005;Fremaux et al. 2006). However, the percentage of stx1-positive isolates was higher in beef cattle (22.73%) than in dairy cattle (3.7%), which was different from the results of a previous study (Bok et al. 2015). Further analysis showed that detection rate of irp2/fyuA combination in E. coli isolates from beef cattle was also higher than that in dairy cattle. Interestingly, irp2/ fyuA/Stx1, hylA/eaeA/Stx1, irp2/fyuA/F41 and irp2/ F41 combinations were not detected in dairy cattle but were detected in beef cattle. These results lay a foundation for further understanding the distribution of virulence genes in E. coli isolated from dairy cattle and beef cattle and provide a basis for reducing E. coli infections.

Conclusions
The results of this study indicated that MDR diarrheagenic E. coli were more common in dairy and beef calves, with frequent MDR, ESBL and the presence of tetracycline resistance gene tet (A). The prevalence rate in dairy cattle is higher than that in beef cattle, which may be related to the prevalence of resistance genes and highlights the importance of the rational use of antimicrobials and strict enforcement of preventive measures in cattle farms. Furthermore, detection rate of virulence genes in the isolates from dairy cattle was lower than that in beef cattle. Although the link between resistance and virulence genes has been extensively studied and virulence genes irp2 and fyuA have a high detection rate in MDR strains, it is still not conclusive. Our results provide important evidences for better exploring their interaction mechanism. Further studies are also needed to understand the origin and transmission route of E. coli in cattle to reduce its prevalence.  (Momtaz et al. 2013a, b).
After overnight incubation at 37°C, only pure pink colonies were selected and transferred to nutrient agar. The isolate was identified by 16S rDNA and stored in 50% glycerol at −80°C.

Statistical analysis
All statistical analyses were performed using GraphPad Prism® 8.00 software (GraphPad Software, Inc., USA). For all experiments, differences were considered to be statistically significant at P< 0.05 values.