Diverse Escherichia coli pathovars of phylogroups B2 and D isolated from animals in Tunisia

Introduction: The virulent Escherichia coli strains responsible for extraintestinal infections were mainly belonged to B2 and D phylogroups. However, no past studies have determinate via the presence of virulence genes the frequency of E. coli pathovars recovered from animals housed in farms in Tunisia. The aims of this study were to investigate 26 E. coli isolated from healthy and diarrheic animals and to determinate via the presence of virulence genes the frequency of pathovars. Methodology: Twenty-six E. coli isolates of phylogroups B2 (n = 14), B22 (n = 9), B23 (n = 5), and D2 (n = 12) were characterized. Genes encoding virulence factors (fimH,eaeA,aggC,papC, papG allele III, hlyA, east1, cnf1, exhA,stx1, stx2, iutA, fyuA, ibeA,and ipaH), and antibiotic resistance as well as class 1 and 2 integrons were searched by polymerase chain reaction (PCR). The genetic relationship of isolates was done by PFGE. Results: According to the occurrence of specific genes the 26 isolates were classified as:9 EAEC, 2 EHEC, 4 UPEC, 3 EPEC/EHEC and 1 NTEC. Therefore, 2 Ex-PEC and 5 APEC were presented amongst our strains. Some isolates (12) were clonal and the remaining was unrelated. Conclusions: Higher diversity of pathovars which carried diverse combinations of virulence genes in healthy isolates. In addition, it seems that the infections were caused by different mechanisms.


Introduction
The majority of Escherichia coli bacterial populations are harmless commensals of mammals [1]. However, in some conditions, they can cause either intestinal or extraintestinal infections. Manifestation of clinical symptomatology and pathology appears to be closely associated with the possession of certain virulence gene combinations that have a range of functions, including toxin production, attachment/invasion, and immune evasion [2][3][4][5].
In poultry farms, another pathotype of ExPEC avian pathogenic E. coli (APEC) strains can cause colibacillosis, which responsible for the mortality of 3%-4% of the animals on a farm, and for a 2%-3% reduction in egg production [11].
Phylogenetic analysis has shown that E. coli comprises four main phylogenetic groups: A, B1, B2, and D. Strains belonging to groups A and B1 are found primarily in the commensal flora. However, pathogenic strains associated with severe acute diarrhea or extraintestinal infections mainly belong to B2 and D phylogroups [12].
E. coli isolated from animals with multiple antibiotic-resistant phenotypes have been reported in Tunisia and worldwide [13,14]. This situation has resulted in a need for more epidemiological information on the prevalence of resistance to various antibiotics and their relevant genes, such as virulence gene combinations in animal isolates.
The aims of this study were to determinate the frequency of the occurrence of potentially pathogenic E. coli strainsbelonging to B2 and D2 phylogroups isolated from healthy and diseased animals in Tunisia, and to detect their virulotypes and their genetic relationship.

Bacterial isolates collection
A total 116 E. coli isolates from healthy and diseased animals (chickens, bovines, and ovines) were recovered from different farms located in nine different governorates in Tunisia between September 2009 and March 2012. Isolates were from poultry feces (n = 61), oral swabs and different organs (n = 13), bovine feces (n = 27), ovine feces (n = 6), and poultry meat (n = 9). Two grams of each fecal sample were homogenized with 2 mL of brain-heart infusion broth, spread onto MacConkey agar plates, and incubated overnight at 37°C.  For organs (trachea, liver, intestine, and heart) and poultry meat, 25 grams were homogenized for 2 minutes with 225 mL of buffered peptone water (Bio-Rad, Marnes la Coquette, France), seeded onto MacConkey agar plates, and incubated for 24 hours at 37°C. Isolates with typical E. coli morphology were selected (one per sample), and the presumptive identification was confirmed by classical biochemical methods and by the API20E system (BioMerieux, Marcy l'Etoile, France).

Statistical analysis
A virulence score was determined for each strain and calculated as the sum of virulence genes detected. Statistical testing was done using EpiInfo software version 6.04 (CDC, Atlanta, USA). Comparisons of proportions were determined using the Chi-squared test or Fisher's exact test.

Detection of class 1 and 2 integrons, phylogenetic groups, and pulsed-field gel electrophoresis (PFGE)
The occurrence of class 1 and class 2 integrons was investigated by PCR [17]. PFGE was performed as described previously by Kaufmann [18], and PFGE profiles were interpretable as recommended by Tenover et al. [19].

Antimicrobial susceptibilities
Among the 26 isolates of E. coli studied, 22 were resistant to tetracycline, 18 to streptomycin, 17 to amoxicillin, 14 to nalidixic acid, 12 to trimethoprim/sulfamethoxazole,7 to ciprofloxacin, and 9 to sulfonamides. The strain resistant to ceftazidime and cefotaxime (EC15) was an extended-spectrum betalactamase (ESBL) producer. No resistance to imipenem or gentamicin was observed. Only 2 isolates were susceptible to all antibiotics, and 18 isolates were multidrug resistant ( Table 2).

Occurrence of integrons and genetic relatedness
Class 1 integrons were found in 18 isolates (Table  2). However, class 2 integrons were not detected. All E. coli isolated from the feces of healthy turkeys showed the same pulsotype (P12). Similarly, the 4 E. coli isolates (EC10; EC11, EC12, and EC13 collected from different organs of 1 chicken with diarrhea were clonally related and belonged to the same pulsotype (P11). However, the remaining strains presented unrelated PFGE patterns.

Statistical analysis
The median virulence score was 3 and ranged from 1 to 6; the ibeA gene was significantly associated with the diarrheic chicken (p = 0.02) and with susceptibility to amoxicillin (p = 0.03).

Discussion
The multidrug resistance trait of E. coli is a cause of concern worldwide. In this study, we found a high level of resistance to tetracycline, streptomycin, amoxicillin, nalidixic acid, trimethoprim-sulfamethoxazole, sulfonamides, and ciprofloxacin. The results also showed low levels of resistance to amoxicillin/clavulanic acid, ceftazidime, and cefotaxime. Similar results have been reported in E. coli strains isolated from animal origins, especially avian isolates, in many countries including Tunisia [20,21]. High rates of antimicrobial resistance in E. coli have been reported in Tunisian patients [22][23][24]. This finding might be linked to the excessive use of antibiotics in clinical settings. However, animal-to-human transmission of resistant E. coli isolates cannot be excluded. Indeed, identical or closely related isolates from humans and animals have been previously reported in the Netherlands, suggesting a likely transmission of E. coli isolates from animals to humans, most probably via the food chain [25].
In our collection, 2 isolates were susceptible to all antibiotics tested, 3 were resistant to 2 families of antibiotics, and 21 isolates were multidrug resistant. It is also interesting to note that all multi-resistant drug isolates were from feces of avian origin, while the other isolates from cows and sheep or meat were resistant just to 2 or 3 families of antibiotics.
Multidrug resistance is mainly linked to integrons. In our study, the presence of class 1 integrons was demonstrated in 18 isolates, while class 2 integrons were detected in only 1 strain. These results are consistent with other studies that showed the dominance of class 1 integrons over class 2 integrons in E. coli of human and animal origin [20,21]. The class 1 integrons were functional and capable of integrating multiple genes cassettes in their variable regions, including their expression, and consequently by providing a common promoter [26].
The 26 strains studied were subdivided into phylogroups B22 (n = 9), B23 (n = 5), and D2 (n = 12). Our selected strains were therefore potentially pathogenic. It is important to note that there is a high risk of pathogenic bacteria spreading to humans via the food chain or through direct contact with farmers and veterinarians, as well as contamination of agricultural soil by animal manure (used as organic fertilizers).
In our study, we looked for 15 different genes encoding virulence factors in E. coli using a PCR technique. The iutA gene was foundin 4 strains without combination with another group of virulence genes such as stx1, stx2, ibeA, eaeA, east1, and cnf1. Therefore, among the 26 E. coli isolates, 4 (15.3%)were UPEC according to the presence of the siderophoreencoding gene iutA,which is responsible for urinary tract infections [9], and 9 (34.6%) were EAEC by the presence of the east1 gene. The UPEC and EAEC pathovars would therefore be most frequently involved in human diarrhea in our environment [27][28]. E. coli is known as the first agent of urinary tract infections [24] in which the UPEC and NTEC pathovars are mostly involved. In our collection, only one NTEC isolate was identified, which was isolated from a healthy cow.
Another group of virulence factors was the adhesins that were considered essential virulence factors in E. coli. These adhesins were encoded by several genes; among them, fimH, eaeA, papC, and papG allele III were decoded in our isolates. The fimH gene was detected in almost of our isolates; this finding is in agreement with the literature, which shows that this gene is the most frequently detected with respect to genes encoding the other adhesins and the rest of the virulence genes [6,29]. In addition, it is usually associated with ExPEC [6].
The ibeA gene was detected in five strains isolated from poultry, which were therefore classified as APEC (19.2%). This virulence factor is known to be involved in crossing the blood-brain barrier in E. coli strains, responsible for neonatal meningitis in humans. It has been reported that the strains harboring this gene are exclusively of avian origin (APEC) [11]. This was confirmed by statistical testing in our collection; we found that the ibeA gene was significantly associated with diarrheic poultry.
PFGE showed that the eight fecal turkey isolates were indistinguishable (PFGE pattern P12). This finding supported intra-transmission of a common clone within this turkey farm, highlighting the wellknown phenomenon of rapid and easy transmission of pathogens within an avian herd. APEC strains cause a wide range of localized and systemic infections commonly called avian colibacillosis, which is one of the leading causes of mortality and morbidity associated with economic losses in the industry throughout the world. In our study, the occurrence of four APEC isolates recovered from different organs of one chicken suffering from diarrhea (P11) supports the systemic form of colibacillosis from a respiratory origin that induces colisepticemia, leading to the dissemination of such strains to different organs.

Conclusions
Our results showed multidrug resistance in the majority of our E. coli isolates, which is in agreement with many reported results of E. coli isolates of animal origins in Tunisia and worldwide. This multidrug resistance trait seems to be linked to the occurrence of class 1 integrons, found in 18 of 26 isolates. Moreover, the occurrence of ibeA and stx1/stx2 genes in some strains is worrisome for human health. The great diversity of pathovars supports the necessity of surveying healthy avian, bovine, and ovine E. coli isolates that could easily be transferred to humans via the food chain, and of successfully identifying risk factors and the major routes of contamination, which determines the control of infections associated with pathovars.